Table data processing method and device, computer equipment, storage medium and program product

Abstract

The invention provides a table data processing method and device, computer equipment, a storage medium and a program product. The method comprises the steps of determining feature information of a target data row in response to an associated data recommendation request for the target data row in a first data table; according to the feature information, obtaining at least one first data row matched with the feature information in a first data table and / or obtaining at least one second data row matched with the feature information in a second data table associated with the first data table; according to at least one first data row and / or at least one second data row matched with the feature information, determining recommended associated data of the target data row; and generating an associated data recommendation result of the target data row according to the recommended associated data. According to the table data processing method and device, the computer equipment, the storage medium and the program product provided by the invention, the coverage rate and accuracy of data recommendation can be improved to a certain extent.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and in particular, to a method and apparatus for processing tabular data, a computer device, a storage medium, and a program product. Background Art

[0002] With the continuous development of Internet technologies, many tasks can be completed intelligently. For example, tables can be created and maintained through collaborative office software.

[0003] In related technologies, a certain field of a certain data row in a data table can be associated with one or more data rows of another data table, so that it is convenient to view information of the other data table associated therewith in this data table.

[0004] However, the inventors of the present disclosure have found that in related technologies, when it is necessary to search for associated records in a table, the operation is relatively cumbersome. Summary of the Invention

[0005] The present disclosure provides a method and apparatus for processing tabular data, a computer device, a storage medium, and a program product to solve or partially solve the above problems.

[0006] In a first aspect of the present disclosure, a method for processing tabular data is provided, including:

[0007] Responding to an associated data recommendation request for a target data row in a first data table, and determining feature information of the target data row;

[0008] Obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in a second data table associated with the first data table according to the feature information;

[0009] Determining recommended associated data of the target data row according to the at least one first data row and / or the at least one second data row that match the feature information;

[0010] Generating an associated data recommendation result for the target data row according to the recommended associated data.

[0011] In a second aspect of the present disclosure, a device for processing tabular data is provided, including:

[0012] A first determination module configured to: respond to an associated data recommendation request for a target data row in a first data table, and determine feature information of the target data row;

[0013] An acquisition module, configured to: obtain at least one first data row that matches the feature information in the first data table and / or obtain at least one second data row that matches the feature information in a second data table associated with the first data table according to the feature information;

[0014] A second determination module, configured to: determine recommended associated data for the target data row according to the at least one first data row and / or the at least one second data row that match the feature information;

[0015] A generation module, configured to: generate an associated data recommendation result for the target data row according to the recommended associated data.

[0016] In a third aspect of the present disclosure, there is provided a computer device, including one or more processors, a memory; and one or more programs, wherein the one or more programs are stored in the memory and executed by the one or more processors, and the one or more programs include instructions for executing the method according to the first aspect.

[0017] In a fourth aspect of the present disclosure, there is provided a non-volatile computer-readable storage medium containing a computer program, which causes the one or more processors to execute the method according to the first aspect when the computer program is executed by the one or more processors.

[0018] In a fifth aspect of the present disclosure, there is provided a computer program product, including one or more computer programs, and the one or more computer programs implement the method as described in the first aspect when executed by one or more processors.

[0019] The table data processing method, device, computer device, storage medium, and program product provided by the embodiments of the present disclosure obtain at least one first data row that matches the feature information in the first data table and / or obtain at least one second data row that matches the feature information in a second data table associated with the first data table according to the feature information of the target data row, and implement automatic recommendation of data based on the at least one first data row and / or the at least one second data row that match the feature information, which is convenient for users to find associated records. Description of the Drawings

[0020] In order to more clearly illustrate the technical solutions in the present disclosure or related technologies, the following will briefly introduce the drawings required for use in the embodiments or related technology descriptions. Obviously, the drawings in the following descriptions are only the embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 Shows a schematic diagram of an exemplary system provided by an embodiment of the present disclosure.

[0022] Figure 2A Shows a schematic diagram of an exemplary page according to an embodiment of the present disclosure.

[0023] Figure 2B Shows another schematic diagram of an exemplary page according to an embodiment of the present disclosure.

[0024] Figure 3A Shows a schematic flowchart of an exemplary table data processing method provided by an embodiment of the present disclosure.

[0025] Figure 3B Shows a schematic flowchart of a method for determining at least one first data row matching feature information according to an embodiment of the present disclosure.

[0026] Figure 3C Shows a schematic flowchart of a method for determining at least one second data row matching feature information according to an embodiment of the present disclosure.

[0027] Figure 3D Shows a schematic flowchart of a method for determining a third candidate data row according to an embodiment of the present disclosure.

[0028] Figure 3E Shows a schematic flowchart of a method for determining a fourth candidate data row according to an embodiment of the present disclosure.

[0029] Figure 3F Shows a schematic flowchart of a method for determining recommended associated data corresponding to a target field of a target data row according to an embodiment of the present disclosure.

[0030] Figure 3G Shows a schematic flowchart of another method for determining recommended associated data corresponding to a target field of a target data row according to an embodiment of the present disclosure.

[0031] Figure 3H Shows a schematic flowchart of yet another method for determining recommended associated data corresponding to a target field of a target data row according to an embodiment of the present disclosure.

[0032] Figure 3I Shows a schematic flowchart of yet another method for determining recommended associated data corresponding to a target field of a target data row according to an embodiment of the present disclosure.

[0033] Figure 4A Shows a schematic diagram of an exemplary page according to an embodiment of the present disclosure.

[0034] Figure 4B Shows another schematic diagram of an exemplary page according to an embodiment of the present disclosure.

[0035] Figure 5 Shows a schematic diagram of the hardware structure of an exemplary computer device provided by an embodiment of the present disclosure.

[0036] Figure 6 Shows a schematic diagram of an exemplary device provided by an embodiment of the present disclosure. Detailed implementation manners

[0037] To make the objectives, technical solutions, and advantages of the present disclosure clearer and more understandable, the following further elaborates on the present disclosure in detail with reference to specific embodiments and the accompanying drawings.

[0038] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present disclosure should have the ordinary meanings understood by those with ordinary skills in the field to which the present disclosure belongs. The "first", "second", and similar terms used in the embodiments of the present disclosure do not represent any order, quantity, or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before the term cover the elements or objects listed after the term and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left", and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0039] It can be understood that before using the technical solutions of the various embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner, and the user's authorization will be obtained.

[0040] For example, when responding to a user's active request, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application program, server, or storage medium that performs the operations of the technical solutions of the present disclosure according to the prompt message.

[0041] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving the user's active request may be, for example, in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry selection controls for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0042] It should be understood that the above-mentioned notification and the process of obtaining user authorization are only illustrative and do not limit the implementation manner of the present disclosure. Other manners that comply with relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0043] Figure 1 FIG. shows a schematic diagram of an exemplary system 100 provided by an embodiment of the present disclosure.

[0044] As Figure 1 shown, the system 100 can be used to implement functions such as creating and maintaining a table, and can include terminal devices 102A, 102B, a server 106, and a database server 108. A medium (e.g., a network) providing a communication link may be included between the terminal devices 102A and the server 106 and the database server 108. The network may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0045] Various application programs (APPs) or software can be installed on the terminal devices 102A, 102B, such as, for example, collaborative office application programs or software, image processing application programs or software, video conferencing application programs or software, reading application programs or software, video application programs or software, social application programs or software, payment application programs or software, web browsers, and instant messaging tools, etc. In some embodiments, these application programs or software can all be used to create and maintain a table, etc.

[0046] The terminal devices 102A, 102B here can be hardware or software. When the terminal devices 102A, 102B are hardware, they can be various electronic devices or computer devices with a display screen, including but not limited to smart phones, tablet computers, e-book readers, MP3 players, laptop computers, and desktop computers (PCs), etc. When the terminal devices 102A, 102B are software, they can be installed in the above-listed electronic devices. It can be implemented as multiple software or software modules (e.g., for providing distributed services), or can be implemented as a single software or software module. No specific limitation is made here.

[0047] The server 106 can be a server providing various services, such as a background server that provides support for various applications or software displayed on the terminal devices 102A, 102B. The database server 108 can also be a database server providing various services. It can be understood that in the case where the server 106 can implement the related functions of the database server 108, the database server 108 may not be provided in the system 100.

[0048] The server 106 and the database server 108 here can also be either hardware or software. When they are hardware, they can be implemented as a distributed server cluster composed of multiple servers or as a single server. When they are software, they can be implemented as multiple software or software modules (such as those used to provide distributed services) or as a single software or software module. Specific limitations are not made here.

[0049] It should be noted that the table data processing method provided by the embodiments of the present disclosure can be executed by the terminal devices 102A and 102B or by the interaction of each device in the system 100. It should be understood that Figure 1 the numbers of the terminal devices, users, servers, and database servers in

[0050] In some exemplary scenarios, the user 104A or the user 104B can respectively create and maintain a table through the collaborative office software or application installed in the terminal device 102A or the terminal device 102B. Optionally, the user 104A and the user 104B can jointly maintain the same table. For example, they can view, add, modify, and delete data rows in the table. Optionally, the table can be a table for multi-dimensional management of a project and can include multiple data tables under multiple dimensions. Therefore, the table can further include multiple data tables. Each data table can be used to store the data corresponding to its corresponding dimension and can have an association relationship with other data tables.

[0051] As an optional embodiment, the user 104A and / or the user 104B can also associate the fields of the data rows in the A data table of the table with the data rows in the B data table, so that the user 104A and / or the user 104B can view the information of the data rows in the B data table associated with the field in the A data table.

[0052] In some embodiments, the association can include one-way association and two-way association.

[0053] One-way association can mean that some fields in the table support the one-way association function, that is, the data rows in the B data table are associated in the A data table. In this way, after the association is completed, the entire row information of the data row can be clicked and viewed in this field of the A data table, and further jump to the B data table associated with this field for viewing. In some embodiments, the fields of one-way association can also be associated with the data rows in the current data table (i.e., the A data table).

[0054] A two-way association may mean that some fields in a table support the two-way association function. By using two-way association fields, one or more data rows in Data Table B can be associated with Data Table A, and the associated data rows in Data Table B will automatically be associated back to the corresponding data rows in Data Table A. Users can directly view the data in Data Table B in Data Table A and can further jump to Data Table B. Moreover, they can also jump back to Data Table A with one click.

[0055] As can be seen, when User 104A and / or User 104B perform an association operation in the current data table (e.g., Data Table A), they need to first find the data rows in another data table (e.g., Data Table B) that need to be associated, and then associate the two through the association operation. However, the inventors of the present disclosure have found that when the number of data rows in another data table (e.g., Data Table B) is very large, users may need to continuously scroll through the data rows of this data table to find the data rows they want to associate, and the operation is not simple enough.

[0056] As mentioned above, User 104A and / or User 104B can use the collaborative office software or application installed in the terminal devices 102A and 102B to create and maintain the table.

[0057] Exemplarily, in the initial state, User 104A can use the terminal device 102A to open the collaborative office software or application, and further open the table that wants to be maintained, or create a new table.

[0058] In some embodiments, the table may include multiple data tables in multiple dimensions. For example, the table may be an order and commodity statistics table, and thus may further include a data table for statistics of orders, a data table for statistics of commodities, and other related data tables, etc.

[0059] Exemplarily, User 104A can further open one of the data tables for data maintenance. For example, in the table, the corresponding label of the data table is triggered to open the data table.

[0060] Figure 2A Shows a schematic diagram of an exemplary page 200 according to an embodiment of the present disclosure.

[0061] As Figure 2A shown, User 104A triggers the first data table, so that the first data table 202 can be displayed in the page 200. The first data table 202 may further include a plurality of first data rows 2022 (exemplarily, Figure 2ATwelve first data rows are shown (2022). Exemplarily, the first data table 202 may be an order table for recording orders. In other words, each data row in the order table records an order. Therefore, the data rows in the data table can also be referred to as records, and the data rows of other data tables associated with the current data table (or its fields) can also be referred to as associated records.

[0062] In some embodiments, as Figure 2A shown, the first data table 202 may further include multiple fields, and each field may correspond to a type of attribute information of the first data row 2022. Taking the order table as an example, it may include multiple fields such as order number, order status, person in charge, business line, order placement time, product library, etc. By filling in the corresponding field information in each first data row 2022, an order can be recorded accordingly.

[0063] Optionally, as Figure 2A shown, the page 200 may further include a field addition control 204, and the user 104A can trigger the field addition control 204 to add a new field to the first data table 202, thereby enriching the content of the first data table 202.

[0064] In some embodiments, when a specific field supports one-way association or two-way association functions, the user 104A can associate the specific field with a data row in another data table. Exemplarily, as Figure 2A shown, the product library field can be a field that can achieve one-way association or two-way association, so that the user 104A can perform an association operation for the product library field in the data row. Optionally, whether a field can be associated with other data tables, which data tables can be associated, and whether this association is one-way association or two-way association can all be pre-configured. Exemplarily, the product library field can be associated with the product table, so that the corresponding data row can be found from the product table and associated to the product library field of the first data table 202.

[0065] Exemplarily, the user 104A can generate a data association request by triggering (for example, clicking or double-clicking) the target field 2024 of the target data row 2022A in the first data table 202 on the page 200 (for example, Figure 2A the third data row in the first data table 202, and this data row can be the data row that the user is currently editing) (for example, Figure 2A the cell corresponding to the product library field of the target data row 2022A), so that the terminal device 102A can further display the first page 210 in response to the data association request, as Figure 2B shown.

[0066] In some embodiments, as Figure 2BAs shown, the first page 210 can be floating above the page 200. Compared with operating by jumping to a page, in this embodiment, it can give the user 104A a feeling of maintaining data for the first data table 202 in the page 200, which is more in line with the user's habits and thus improves the user experience.

[0067] As Figure 2B shown, optionally, the second data table 212 associated with the first data table 202 can be further displayed in the first page 210, so that the user 104A can search for the second data rows 2122 of the second data table 212 that he wants to associate in the first page 210. Optionally, the second data table 212 is also associated with the target field 2024, and the association relationship between the second data table 212 and the target field 2024 can be pre-configured on the corresponding field of the target field 2024. Thus, when the target field 2024 is triggered, the content of the second data table 212 can be displayed on the first page 210. Exemplarily, the second data table 212 can be a product table for recording product information, and the associated one can be the product library field of the first data table 202.

[0068] As Figure 2B shown, the second data table 212 can include multiple second data rows 2122. When the display range of the first page 210 is not sufficient to display all the second data rows 2122 of the second data table 212, in some embodiments, a scroll bar 2124 can be displayed on one side of the second data table 212. The scroll bar 2124 is used to scroll the second data table 212 in the first direction (for example, the vertical direction of the first page 210), so that the user 104A can slide to view other second data rows 2122 in the second data table 212 by moving the scroll bar 2124, and the operation is relatively convenient.

[0069] In some embodiments, as Figure 2B shown, the first page 210 can also include a search bar 214, so that the user 104A can search for the second data rows 2122 that he wants to associate in the second data table 212 by entering a search term in the search bar 214, thereby facilitating the user to find the second data rows 2122 that he wants to associate more quickly. Optionally, the search bar 214 can be displayed at the top of the first page 210, so that the user 104A can more easily observe the search bar 214 and thus use it more conveniently.

[0070] However, the inventors of the present disclosure have found that when the number of second data rows 2122 in the second data table 212 is large, the first page 210 cannot fully display all the content of the second data table 212. In particular, when the number of second data rows 2122 in the second data table 212 is extremely large, the user 104A may have difficulty finding the second data row that they want to associate. Even when using the search bar 214 to conduct a search, it may be impossible to obtain a suitable result due to inaccurate search terms.

[0071] In some embodiments, methods such as keyword retrieval, vector similarity calculation, and data analysis can be used to recommend data rows that can be associated to the user. However, the inventors of the present disclosure have found that there are certain limitations in using any one of the above methods (i.e., using a single recall strategy) for recommendation. For example, problems such as insufficient coverage, low accuracy, and lack of personalization exist. Specifically, a single recall strategy is difficult to cover all possible association scenarios and may miss some valuable association records; a single recall strategy may be affected by factors such as data sparsity, semantic ambiguity, and noisy data, resulting in inaccurate recommendation results; a single recall strategy is difficult to fully consider the user's personalized needs and scenario differences, and the recommendation results are relatively general but lack pertinence.

[0072] In view of this, embodiments of the present disclosure provide a method for processing tabular data. By integrating at least two recall strategies, the coverage rate and accuracy of data recommendation can be improved to a certain extent, thereby solving or partially solving the above problems to a certain extent.

[0073] Figure 3A The flowchart of an exemplary tabular data processing method 300 provided by embodiments of the present disclosure is shown. This method 300 can be used to recommend data rows that can be associated. Optionally, this method 300 can be independently implemented by Figure 1 the terminal device 102A or 102B, or, by Figure 1 the server 106 independently, or can also be interactively implemented by Figure 1 each device in the Figure 3A system 100. As

[0074] shown, this method 300 may further include the following steps.

[0075] Optionally, the triggering of the associated data recommendation request can be triggered by clicking on the recommendation control 216 of the first page 210, so that the terminal device 102A can know that the user 104A currently needs to use the function of the associated data of the recommended target data row 2022A. In some embodiments, the first page 210 is displayed by triggering the target field 2024 of the target data row 2022A. Therefore, the recommendation control 216 can be used to recommend associated data for the target field 2024 of the target data row 2022A. Thus, the associated data recommendation request can be an associated data recommendation request for data recommendation for the target field 2024 of the target data row 2022A. In this way, the obtained recommended associated data can be used to be associated with the target field 2024 of the target data row 2022A. In some embodiments, as Figure 2A shown, when the user 104A triggers the target data row 2022A (e.g., clicks or double-clicks on the target data row 2022A) in the Figure 2A page 200, the associated data recommendation request can also be generated, so that the recommended associated data can be used to be associated with the target data row 2022A (e.g., after obtaining the recommended associated data, by selecting the corresponding field to associate the associated data with the field).

[0076] In this step, the feature information of the target data row 2022A can be used to express various features of the target data row 2022A, so that based on this feature information, the data rows associated with the target data row 2022A can be obtained. It can be understood that the feature information can be the feature information obtained from the target data row 2022A by any method of extracting features, as long as it can ensure that the feature information can express the characteristics of the target data row 2022A.

[0077] In some embodiments, the determining of the feature information of the target data row may further include: extracting a plurality of keywords from the target data row, and determining the feature information according to the plurality of keywords. In this way, by extracting a plurality of keywords of the target data row, it can be used to characterize the characteristics of the target data row 2022A, and thus can be used as the feature information of the target data row 2022A.

[0078] It can be understood that the method for extracting keywords can be various methods applicable to keyword extraction. In some embodiments, the target data row 2022A can be input into a keyword extraction model, and the keyword extraction model outputs multiple keywords for the target data row 2022A. The keyword extraction model can be a machine learning model (e.g., a neural network model), and the model can be trained using the data rows of the existing data table, such that the keyword extraction model can extract multiple keywords from the target data row 2022A that can better express the characteristics of the target data row 2022A, thereby improving the subsequent data recommendation effect.

[0079] In some embodiments, the determining the feature information according to the multiple keywords may further include: generating a feature vector corresponding to the target data row according to the multiple keywords, and then determining the feature information according to the multiple keywords and the feature vector.

[0080] In this embodiment, after obtaining multiple keywords, the multiple keywords can also be converted into a feature vector based on vectorization technology, and then both the multiple keywords and the feature vector are determined as the feature information of the target data row 2022A. In this way, making recommendations based on two types of feature information can improve the coverage rate and / or accuracy of the data.

[0081] It can be understood that the vectorization technology can be various methods applicable to generating a feature vector based on keywords. Optionally, the vectorization technology can be a one-hot encoding algorithm, a vocabulary mapping algorithm (e.g., Word2Vec algorithm), a word embedding algorithm, and so on.

[0082] In some embodiments, instead of generating the feature vector according to multiple keywords, the feature vector can also be directly generated based on the target data row 2022A, so that the feature vector can directly represent the feature information of the target data row 2022A without being affected by the keyword extraction algorithm, thereby improving the coverage rate and accuracy of the recommended data.

[0083] After determining the feature information, in step 304, at least one first data row 2022 that matches the feature information can be obtained from the first data table 202 and / or at least one second data row 2122 that matches the feature information can be obtained from the second data table 212 associated with the first data table 202 according to the feature information.

[0084] In this step, after obtaining the feature information of the target data row 2022A, it is possible to perform a match in the first data table 202 based on this feature information to obtain the first data row 2022 that matches this feature information, and / or it is also possible to perform a match in the second data table 212 based on this feature information to obtain the second data row 2122 that matches this feature information. In this way, by using the feature information to perform feature matching in the first data table 202 or the second data table 212 to obtain the recommended first data row 2022 or the second data row 2122, it is possible to achieve the automatic recommendation of associated data (or associated records), thereby facilitating the user to find the data row they want to associate in the recommendation results.

[0085] In some embodiments, it is possible to simultaneously perform a match in the first data table 202 based on the feature information to obtain the first data row 2022 that matches this feature information and perform a match in the second data table 212 to obtain the second data row 2122 that matches this feature information, so that two recall strategies can be used to obtain the associated data for recommendation. Compared with using a single recall strategy, the coverage rate and accuracy of the recommended data can be improved.

[0086] Specifically, since there are several first data rows 2022 in the first data table 202, and these first data rows 2022 similar to the target data row 2022A are also associated with the second data row 2122, then taking the second data row 2122 associated with the first data row 2022 similar to the target data row 2022A as the recommended data can transfer the existing association experience to the target data row 2022A, thereby expanding the coverage range of the recommended data and improving the recall rate.

[0087] In some embodiments, as Figure 3B shown, the step of obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in the second data table associated with the first data table according to the feature information may further include the following steps:

[0088] In step 3042, according to the multiple keywords, obtain at least one first candidate data row that matches the multiple keywords in the first data table.

[0089] In this step, based on the multiple keywords of the target data row 2022A, at least one first data row 2022 that matches the multiple keywords can be found in the database corresponding to the first data table 202 to be used as the at least one first candidate data row corresponding to the target data row 2022A. Optionally, the first data row 2022 with a matching degree higher than the matching degree threshold can be used as the first candidate data row.

[0090] It can be understood that any keyword matching algorithm can be used to find the first candidate data row. Optionally, a string matching algorithm (e.g., the KMP (Knuth-Morris-Pratt) algorithm), a trie tree algorithm, a multi-pattern matching algorithm (e.g., the Aho-Corasick automaton algorithm), etc. can be adopted.

[0091] In step 3044, according to the feature vector, at least one second candidate data row that matches the feature vector is obtained in the first data table.

[0092] In this step, based on the feature vector of the target data row 2022A, at least one first data row 2022 that matches the feature vector can be found in the vector database corresponding to the first data table 202 as the at least one second candidate data row corresponding to the target data row 2022A.

[0093] Optionally, calculating the cosine similarity of vectors can be used as the method for calculating similarity. Optionally, the first data row 2022 with a similarity higher than the similarity threshold can be used as the second candidate data row.

[0094] In step 3046, according to the at least one first candidate data row and the at least one second candidate data row, at least one first data row that matches the feature information is determined.

[0095] In this step, the at least one first candidate data row and the at least one second candidate data row can be combined into a data set as the at least one first data row that matches the feature information, so as to expand the range of the at least one first data row that matches the feature information and improve the data coverage.

[0096] In some embodiments, the intersection can also be taken from the at least one first candidate data row and the at least one second candidate data row as the at least one first data row that matches the feature information. The first data row obtained in this way satisfies both the keyword matching degree requirement and the vector similarity requirement, which can improve the data accuracy.

[0097] Moreover, since the filled information of the target data row 2022A can contain rich text information and keyword fields, keyword extraction can be supported, which is beneficial to obtaining recommended data through keyword matching and vector similarity calculation subsequently.

[0098] In some embodiments, such as Figure 3CAs shown, obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in a second data table associated with the first data table according to the feature information may further include the following steps:

[0099] In step 3048, according to the multiple keywords, obtain at least one third candidate data row that matches the multiple keywords in the second data table.

[0100] In this step, based on the multiple keywords of the target data row 2022A, at least one second data row 2122 that matches the multiple keywords can be found in the database corresponding to the second data table 212 as the at least one third candidate data row corresponding to the target data row 2022A. Optionally, the second data row 2122 with a matching degree higher than the matching degree threshold can be used as the third candidate data row.

[0101] It can be understood that any keyword matching algorithm can be used to find the third candidate data row. Optionally, a string matching algorithm (e.g., the KMP (Knuth-Morris-Pratt) algorithm), a trie tree algorithm, a multi-pattern matching algorithm (e.g., the Aho-Corasick automaton algorithm), etc. can be adopted.

[0102] In step 3050, according to the feature vector, obtain at least one fourth candidate data row that matches the feature vector in the second data table.

[0103] In this step, based on the feature vector of the target data row 2022A, at least one second data row 2122 that matches the feature vector can be found in the vector database corresponding to the second data table 212 as the at least one fourth candidate data row corresponding to the target data row 2022A.

[0104] Optionally, calculating the cosine similarity of vectors can be used as the method for calculating similarity. Optionally, the second data row 2122 with a similarity higher than the similarity threshold can be used as the fourth candidate data row.

[0105] In step 3052, determine the at least one second data row that matches the feature information according to the at least one third candidate data row and the at least one fourth candidate data row.

[0106] In this step, the at least one third candidate data row and the at least one fourth candidate data row can be merged into a data set as the at least one second data row that matches the feature information, so as to expand the range of the at least one second data row that matches the feature information and improve the data coverage rate.

[0107] In some embodiments, an intersection can also be taken from the at least one third candidate data row and the at least one fourth candidate data row as the at least one second data row that matches the feature information. The obtained second data row meets both the keyword matching degree requirement and the vector similarity requirement, which can improve the data accuracy.

[0108] In addition to the foregoing recall strategies, in some embodiments, as Figure 3D shown, the method 300 also provides a recall strategy based on the most recent edit records, and may further include the following steps:

[0109] In step 3054, determine at least one first edited data row generated within a first preset time period (e.g., within 3 days, within 5 days, within one week, etc.) from the current time in the first data table 202.

[0110] In this step, at the current time (or current moment), the first data rows in the first data table 202 that have been edited within the first preset time period (e.g., within 3 days, within 5 days, within one week, etc.) can be determined as the first edited data rows.

[0111] In step 3056, determine at least one data row in the second data table 212 associated with the target field (e.g., the product library field) of the at least one first edited data row as at least one third candidate data row.

[0112] In this step, after finding the at least one first edited data row, it can be determined whether the target field (e.g., the product library field) of these first edited data rows is associated with a certain second data row 2122 in the second data table 212. If the target field of these first edited data rows is associated with a certain or some second data rows 2122 in the second data table 212, then the certain or some second data rows 2122 can be used as the third candidate data row.

[0113] In this embodiment, since the data rows recently edited by the user have a certain similarity in theme or requirements with the current target data row 2022A and may reflect recently popular or active records, thus, making recommendations based on at least one third candidate data row can further improve the data coverage rate of the recommendations and enhance the recall rate.

[0114] In some embodiments, as Figure 3E shown, the method 300 further provides a recall strategy based on similar associated records and may further include the following steps:

[0115] In step 3058, in response to the third data row being associated in the target field of the target data row 2022A, determine at least one other data row in the first data table 202 that associates the third data row in the target field.

[0116] In this embodiment, the target field can be a field that can associate multiple records, and if a record (i.e., the third data row) has been associated in the target field of the target data row 2022A, then other data rows that are the same as the target data row 2022A in associating this record (i.e., the third data row) in the target field can be found in the first data table 202.

[0117] It can be understood that the data table associated with the target field is the second data table 212. Therefore, the data row already associated in the target field of the target data row 2022A also belongs to the second data table 212. Here, to distinguish it from the second data row in name, the data row corresponding to the third data row is named the third data row. In fact, the third data row can also be the second data row 2122 in the second data table 212.

[0118] For clearer illustration, taking the second data table 212 as a product table as an example, the third data row can be the second data row corresponding to product A in the second data table 212. In this step, first data rows that also associate the second data row corresponding to product A can be searched for in the first data table 202 as the other data rows.

[0119] In step 3060, in response to the number of the at least one other data row being greater than a first number (e.g., 2, 3, 4, 5, 10, etc.), determine at least one fourth data row that the at least one other data row also associates in the target field.

[0120] Still taking the second data table 212 as the product table as an example, in this step, when the number of the first data rows associated with the second data row corresponding to product A is greater than the first number, it indicates that the association records of the second data row corresponding to product A appear more frequently in the first data table 202. Then, it can be further determined whether at least one fourth data row (for example, the second data row corresponding to product B) is also associated with among these other data rows associated with the second data row corresponding to product A. Here, in order to distinguish from the second data row in terms of name, the data row corresponding to the fourth data row is named the fourth data row. In fact, the fourth data row can also be the second data row 2122 in the second data table 212.

[0121] In step 3062, for each of the fourth data rows, determine whether the number of the at least one other data row associated with the fourth data row is greater than a second number (for example, 2, 3, 4, 5, 10, etc.).

[0122] Still taking the second data table 212 as the product table as an example, in this step, it can be determined whether the number of the other data rows associated with the second data row corresponding to product B is greater than the second number to determine whether the second data row corresponding to product B appears frequently among these other data rows.

[0123] In step 3064, determine at least one of the fourth data rows with the number of the associated at least one other data row greater than the second number as at least one fourth candidate data row.

[0124] Still taking the second data table 212 as the product table as an example, in this step, when the number of the other data rows associated with the second data row corresponding to product B is greater than the second number, it is determined that the second data row corresponding to product B appears frequently among these other data rows, and thus it can be used to recommend to the user.

[0125] In this embodiment, several first data rows 2022 of the first data table 202 have selected multiple association records (associated with multiple second data rows), which can accumulate sufficient co-occurrence records, form reliable association rules, and improve the recall rate while also improving the recommendation accuracy.

[0126] Back to Figure 3A , after obtaining some candidate data rows, in step 306, the recommended associated data of the target data row can be determined according to the at least one first data row and / or the at least one second data row matching the feature information.

[0127] Optionally, after obtaining the at least one first data row and / or the at least one second data row matching the feature information, the recommended associated data of the target data row 2022A can be determined based on these data rows.

[0128] In some embodiments, the associated data recommendation request includes a request for recommending associated data for a target field of the target data row;

[0129] As Figure 3F shown, determining the recommended associated data of the target data row according to the at least one first data row and / or the at least one second data row matching the feature information may further include the following steps:

[0130] In step 3066, at least one data row in the second data table associated with the target field of the at least one first data row matching the feature information is determined as at least one first candidate data row.

[0131] In this step, the data rows in the second data table associated with the target field of the at least one first data row matching the feature information can be found, and then the data row is used as the first candidate data row.

[0132] For clearer illustration, still taking the commodity table as an example, assuming that multiple first data rows matching the feature information are found in the first data table 202, then the target field (for example, the commodity library field) can be found to be associated with a certain or some second data rows in the commodity table, and then these second data rows with associated records are used as the first candidate data rows.

[0133] In step 3068, the at least one second data row matching the feature information is determined as at least one second candidate data row.

[0134] In this step, the at least one second data row found in the second data table 212 and matching the feature information can be directly determined as the at least one second candidate data row.

[0135] In step 3070, according to the at least one first candidate data row and the at least one second candidate data row, the recommended associated data corresponding to the target field of the target data row is determined.

[0136] In this step, if the at least one first candidate data row and the at least one second candidate data row are both the second data rows 2122 in the second data table 212, then based on them, the data available for association can be recommended for the target field of the target data row 2022A. The data to be recommended incorporates two recall strategies, which can improve the coverage of the recommended data.

[0137] As described above, in some embodiments, at least one third candidate data row may also be obtained based on a recall policy of recent edit records. Therefore, the step of determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row and the at least one second candidate data row may further include: determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, and the at least one third candidate data row. That is, regarding the at least one third candidate data row obtained based on the recall policy of recent edit records as also the data to be recommended can further improve the coverage of the recommended data.

[0138] In some embodiments, at least one fourth candidate data row may also be obtained based on a recall policy of similar associated records. Therefore, the step of determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, and the at least one third candidate data row may further include the following steps:

[0139] Determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, the at least one third candidate data row, and the at least one fourth candidate data row. That is, regarding the at least one fourth candidate data row obtained based on the recall policy of similar associated records as also the data to be recommended can further improve the coverage and accuracy of the recommended data.

[0140] It can be understood that the four recall policies in the embodiments of the present disclosure can be arranged and combined to obtain multiple fusion recall policies, and these fusion policies can all be used as the recall policies in the embodiments of the present disclosure to obtain the data to be recommended. Compared with a single recall policy, a higher recall rate can be obtained, the data coverage can be improved, and a wider range of application scenarios can be applied.

[0141] In some embodiments, the at least one first candidate data row and the at least one second candidate data row are data obtained based on feature matching. To further improve the data accuracy, the two can be screened.

[0142] As Figure 3G shown, the step of determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, the at least one third candidate data row, and the at least one fourth candidate data row may further include the following steps:

[0143] In step 30702, input the at least one first candidate data row into a first screening model to obtain at least one first screened data row.

[0144] Optionally, the first screening model is a large language model (LLM). The first screening model screens the at least one first candidate data row based on a first prompt, and the first prompt includes task information for screening the at least one first candidate data row based on the association relationship between the first candidate data row and the first data table.

[0145] Since the first candidate data row is obtained based on the first data row 2022 screened from the first data table 202, when screening the first candidate data row, it is necessary to consider the association relationship between the first candidate data row and the first data table. In this embodiment, by adding, in the first prompt, task information for screening the at least one first candidate data row based on the association relationship between the first candidate data row and the first data table, the first screening model can screen the at least one first candidate data row based on the association relationship between the first candidate data row and the first data table, so as to obtain more accurate data to be recommended, that is, the at least one first screened data row.

[0146] In step 30704, input the at least one second candidate data row into a second screening model to obtain at least one second screened data row.

[0147] Wherein, the second screening model is a large language model. The second screening model screens the at least one second candidate data row based on a second prompt, and the second prompt includes task information for screening the at least one second candidate data row based on the association relationship between the first data table and the second data table.

[0148] Since the first candidate data row is obtained based on the second data row 2122 screened from the second data table 212 according to the feature information of the target data row 2022A, when screening the second candidate data row, it is necessary to consider the association relationship between the first data table and the second data table. In this embodiment, by adding, in the second prompt, task information for screening the at least one second candidate data row based on the association relationship between the first data table and the second data table, the second screening model can screen the at least one second candidate data row based on the association relationship between the first data table 202 and the second data table 212, so as to obtain more accurate data to be recommended, that is, the at least one second screened data row.

[0149] In step 30706, based on the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row, determine the recommended associated data corresponding to the target field of the target data row.

[0150] In this embodiment, by filtering the first candidate data row and the second candidate data row, more accurate data to be recommended can be obtained.

[0151] In some embodiments, as Figure 3H shown, the determining of the recommended associated data corresponding to the target field of the target data row based on the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row may further include the following steps:

[0152] In step 30708, determine at least one duplicate data row (i.e., the same data row that appears repeatedly) among the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row.

[0153] In step 30710, retain one of the at least one duplicate data row and delete the remaining duplicate data rows.

[0154] In this step, for each group of the same data rows that appear repeatedly, retain one of them, and then delete the remaining data rows. In other words, for each group of the same data rows that appear repeatedly, only retain one as a representative.

[0155] It can be understood that there may be multiple groups of the same data rows that appear repeatedly. For each group of the same data rows that appear repeatedly, the aforementioned processing method is adopted, that is, only retain one as a representative.

[0156] Optionally, after the deduplication process, information such as the multi-path identifier of the retained duplicate data row (i.e., the identifier of other deleted duplicate data rows), the occurrence frequency of this group of duplicate data rows, and the most recent update timestamp of this group of duplicate data rows can also be recorded, so that these information can be used for other processing subsequently.

[0157] In step 30712, based on the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row after deleting the remaining duplicate data rows, determine the recommended associated data corresponding to the target field of the target data row.

[0158] Through this embodiment, duplicate data can be removed from the data to be recommended, ensuring that no identical data rows appear in the final recommendation results and improving the user experience.

[0159] In some embodiments, as Figure 3I shown, determining the recommended associated data corresponding to the target field of the target data row according to the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row may further include the following steps:

[0160] In step 30714, sort the at least one first screened data row according to the occurrence times of identical data rows in the at least one first screened data row and / or the edit time of each first screened data row to obtain a first internal sorting corresponding to the at least one first screened data row.

[0161] Since the first screened data row is obtained by first matching the first data row 2022 in the first data table 202 based on feature information and then finding the second data row 2122 associated with the first data row 2022, it can be understood that the second data rows 2122 associated with the first data row 2022 obtained by matching based on feature information in the first data table 202 may be the same. Therefore, in this step, sorting can be performed according to the occurrence times of identical data rows in the at least one first screened data row.

[0162] For example, still taking the product table as an example, one of the at least one first screened data rows is the second data row corresponding to product A, and this second data row actually appears 5 times before duplicate removal. Another of the at least one first screened data rows is the second data row corresponding to product B, and this second data row actually appears 3 times before duplicate removal. In this way, the second data row corresponding to product A can be ranked in front of the second data row corresponding to product B and be recommended preferentially.

[0163] In some embodiments, sorting can also be performed in combination with the latest edit time. For example, for multiple first screened data rows with the same occurrence times, they can be sorted according to the time of being edited in order of time proximity, that is, the most recently edited data row is ranked in front.

[0164] In this way, the at least one first screened data row is internally sorted according to the first internal sorting, so that the data rows that better meet the user's needs can be recommended to the user preferentially.

[0165] In step 30716, sort the at least one second screened data row according to the matching degree with the multiple keywords and / or the similarity with the feature vector corresponding to the at least one second screened data row, so as to obtain the second internal sorting corresponding to the at least one second screened data row.

[0166] Since the second screened data row is obtained by matching the second data row 2122 in the second data table 212 based on the feature information (keywords and / or feature vectors), therefore, sorting can be performed based on the matching degree with the multiple keywords and / or the similarity with the feature vector corresponding to the second screened data row.

[0167] For example, sorting can be first performed based on the keyword matching degree, and then for those with the same keyword matching degree, sorting can be further performed based on the feature vector similarity. Or, conversely, sorting can be first performed based on the feature vector similarity, and then for those with the same feature vector similarity, sorting can be further performed based on the keyword matching degree.

[0168] In this way, the at least one second screened data row is internally sorted according to the second internal sorting, so that the data rows that better meet the user's needs can be preferentially recommended to the user.

[0169] In step 30718, sort the at least one third candidate data row according to the occurrence times of the same data row in the at least one third candidate data row and / or the editing time of each third candidate data row, so as to obtain the third internal sorting corresponding to the at least one third candidate data row.

[0170] Since the third candidate data row is obtained by first selecting at least one edited data row from the first data table 202 and then determining the second data row associated with the target field of the edited data row, it can be understood that the second data rows 2122 associated with these edited data rows may be the same. Therefore, in this step, sorting can be performed according to the occurrence times of the same data row in the at least one third candidate data row.

[0171] For example, still taking the commodity table as an example, one of the at least one third candidate data rows is the second data row corresponding to commodity A, and this second data row actually appears 5 times before deduplication processing. Another one of the at least one third candidate data rows is the second data row corresponding to commodity B, and this second data row actually appears 3 times before deduplication processing. In this way, the second data row corresponding to commodity A can be ranked in front of the second data row corresponding to commodity B and be preferentially recommended.

[0172] In some embodiments, sorting can also be performed in combination with the latest editing time. For example, for multiple third candidate data rows with the same occurrence count, they can be sorted based on the editing time in ascending order of time, that is, the latest edited data row is ranked first.

[0173] In this way, the at least one third candidate data row is internally sorted according to the third internal sorting, so that the data rows that better meet the user's needs can be preferentially recommended to the user.

[0174] In step 30720, according to the occurrence probability of each of the at least one fourth candidate data row among the at least one fourth candidate data row, the at least one fourth candidate data row is sorted to obtain the fourth internal sorting corresponding to the at least one fourth candidate data row.

[0175] Since the fourth candidate data row is obtained based on the recall strategy with co-occurrence records with the target data row 2022A in the first data table 202, therefore, the conditional probability of the fourth candidate data row relative to the target data row 2022A can be calculated, and then sorted based on the conditional probability.

[0176] For example, still taking the product table as an example, if the target data row 2022A has already associated the second data row corresponding to product A, and one of the at least one fourth candidate data row is the second data row corresponding to product B, then the conditional probability of also associating product B under the condition of associating product A can be calculated, so as to determine the conditional probability of the second data row corresponding to product B. According to this method, the conditional probability of each fourth candidate data row can be calculated, and then sorted based on this.

[0177] In this way, the at least one fourth candidate data row is internally sorted according to the fourth internal sorting, so that the data rows that better meet the user's needs can be preferentially recommended to the user.

[0178] In step 30722, determine the external sorting among the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row.

[0179] In this step, the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row can also be sorted by priority (i.e., external sorting). That is, taking the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row as four sets, set priorities between the sets, so that after merging these four sets, it can be determined which set of data rows can be preferentially recommended.

[0180] In some embodiments, the external sorting is in the order of the fourth internal sorting, the second internal sorting, the first internal sorting, and the third internal sorting, that is, the at least one fourth candidate data row is preferentially recommended (and recommended in sequence according to its fourth internal sorting), followed by the at least one second screened data row, then the at least one first screened data row, and finally the at least one third candidate data row.

[0181] In this embodiment, since the fourth candidate data row is recommended based on co-occurrence association records, it can better meet user needs. Followed by the second screened data row directly matched in the second data table 212 based on feature information, which can also better meet user needs. Then there is the corresponding second data row found after obtaining the first data row in the first data table 202 based on similar rows, which can also reflect user needs to a certain extent. Finally, there is the corresponding second data row found after obtaining the first data row in the first data table 202 based on the most recently edited, which has a lower possibility of meeting user needs but can still play a certain role in recommendation.

[0182] In step 30724, based on the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row are sorted to obtain the recommended associated data.

[0183] In this embodiment, the recommended associated data is sorted using a special sorting rule, so that the data rows that are more likely to meet user needs can be ranked in the front for preferential recommendation, thereby improving the recommendation effect.

[0184] It can be understood that the above embodiments are described in terms of being able to obtain the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row. In fact, each of the above recall strategies can have certain activation conditions, so that the recalled data can better meet the requirements or have a higher recall rate.

[0185] Therefore, in some embodiments, according to the feature information, obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in the second data table associated with the first data table may further include the following steps:

[0186] In response to the target data row 2022A not being empty (i.e., feature information can be extracted from the target data row 2022A) and the first data table 202 including the first data row 2022 for which the association operation has been performed (i.e., the first data table 202 includes the first data row 2022 with an existing association record), at least one first data row matching the feature information can be obtained from the first data table according to the feature information.

[0187] In this way, starting the recall policy only when the start condition is met can improve the recall rate.

[0188] In some embodiments, obtaining at least one first data row matching the feature information from the first data table according to the feature information and / or obtaining at least one second data row matching the feature information from a second data table associated with the first data table may further include the following steps:

[0189] In response to the target data row 2022A not being empty (i.e., feature information can be extracted from the target data row 2022A) and the number of second data rows 2122 in the second data table 212 being greater than a fourth quantity (e.g., 5000, 10000, etc.), at least one second data row 2022 matching the feature information is obtained from the second data table 212 associated with the first data table 202 according to the feature information.

[0190] In this way, when the number of second data rows 2122 in the second data table 212 is large, adopting this recall policy can obtain data rows that better meet user needs and also improve the recall rate.

[0191] In some embodiments, determining at least one first edited data row generated within a first preset time period from the current time in the first data table may further include:

[0192] In response to the first data table including at least one first edited data row generated within the first preset time period from the current time and the target field of the first edited data row having had the association operation performed (i.e., its target field is associated with a second data row), the at least one first edited data row is obtained.

[0193] In this way, starting the recall policy only when the start condition is met can improve the recall rate.

[0194] In some embodiments, in response to a third data row being associated in the target field of the target data row, determining at least one other data row in the first data table that associates the third data row in the target field may further include:

[0195] In response to determining that the target field allows associating multiple second data rows (i.e., this field allows adding multiple records), the proportion of the first data rows in the first data table where the target field has associated multiple second data rows is higher than a preset ratio (for example, the number of first data rows with multiple second data rows associated in the target field accounts for 20%, 30%, 50%, etc. of the total number of data rows in the first data table) and the number of the first data rows in the first data table is greater than a fifth quantity (for example, the total number of data rows in the first data table is 20, 50, etc.), when there is a third data row associated in the target field of the target data row, determine at least one other data row in the first data table that associates the third data row in the target field.

[0196] In this way, the recall strategy is only activated when the activation conditions are met, enabling co-occurrence records to better meet the user's needs.

[0197] It can be understood that it is possible that all the above recall strategies are triggered with a very small probability but still no data can be recalled. Therefore, in some embodiments, a fallback recall strategy is also set. Specifically, the method may further include the following steps:

[0198] In response to failing to obtain the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row, determine at least one second edited data row within a second preset time period (for example, 1 day, 2 days, 3 days, 5 days, etc.) from the current time in the second data table 212 or the third edited data row of the sixth quantity (for example, the nearest 20) from the current time as the recommended associated data of the target data row.

[0199] In this way, by using the most recently edited second data row 2122 in the second data table 212 as the fallback recommended data, the normal execution of the recommendation function can be ensured.

[0200] Back to Figure 3A , after obtaining the recommended associated data, in step 308, according to the recommended associated data, generate an associated data recommendation result for the target data row 2022A.

[0201] In this step, a part can be selected from the recommended associated data as the associated data recommendation result for the target data row 2022A. Since the recommended associated data can be obtained based on a fused recall strategy, a specific screening strategy, and / or a sorting strategy, the associated data recommendation result generated based on it can better meet the user's needs.

[0202] In some embodiments, generating the associated data recommendation result of the target data row according to the recommended associated data may further include: based on the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, selecting a third number (e.g., 3, 5, etc.) of data rows from the recommended associated data as the associated data recommendation result of the target data row.

[0203] In this embodiment, by selecting a third number of data rows from the recommended associated data for recommendation according to the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, the data rows with higher rankings can be preferentially recommended to the user, thus better meeting the user's needs.

[0204] Figure 4A FIG. shows a schematic diagram of an exemplary page 200 according to an embodiment of the present disclosure.

[0205] As Figure 4A shown, in some embodiments, when the user 104A triggers the recommendation control 216, the associated data recommendation result including a plurality of recommended data rows 220 may be displayed in the first page 210.

[0206] As Figure 4A shown, optionally, the first page 210 may include a replacement control 222. After the user 104A triggers the replacement control 222, the terminal device 102A may sequentially select another third number of data rows from the recommended associated data to replace the currently displayed plurality of recommended data rows 220, thereby completing the replacement of the data rows, enabling the user 104A to more easily find the data rows they want to associate.

[0207] As Figure 4A shown, in some embodiments, a check box 2202 is included at the forefront of each data row. By the user 104A checking the corresponding check box 2202 and clicking the confirmation control 250, the data row corresponding to the check box 2202 can be associated with the target data row 2202A, thereby completing the data association. Optionally, as Figure 4A shown, when the user 104A checks the check box 2202 in the recommended data row 220, the check box of the corresponding second data row 2122 in the second data table 212 is also selected, thereby prompting the user 104A that the check operation is for a certain second data row 2122 in the second data table 212, maintaining the consistency between the recommended data row and the second data row.

[0208] Figure 4B FIG. shows another schematic diagram of an exemplary page 200 according to an embodiment of the present disclosure.

[0209] In some embodiments, after user 104A selects a certain recommended data row from the associated data recommendation results for an association operation (e.g., user 104A checks the checkbox 2202 corresponding to a specific second data row and clicks the OK control 250), as Figure 4B shown, the second data row is associated with the target data row 2022A in the first data table 202, and the information of the second data row is displayed in the cell 2024A corresponding to the target field 2024 in the target data row 2022A. Still taking the second data table 212 as the product table as an example, as Figure 4A shown, when user 104A selects the first recommended data row in the associated data recommendation results, the product corresponding to this recommended data row is a necklace. At this time, as Figure 4B shown, the name of this product (e.g., necklace) can be displayed in the cell 2024A corresponding to the target field 2024 in the target data row 2022A, so that user 104A can know that the necklace product in the product table has been associated with the target data row 2022A. It can be understood that the information displayed in the cell 2024A can also be other information that can characterize the content or characteristics of the associated second data row. Taking the second data table 212 as the product table as an example, the information displayed in the cell 2024A can also be other information that can characterize the product, such as the product identifier (ID).

[0210] In some embodiments, when the above association operation is a one-way association, user 104A can also trigger the cell 2024A of the target data row 2022A to view the information of the second data row associated therewith, so as to facilitate the user to view the detailed information of the data row in other data tables associated with the target data row. For example, by hovering the mouse over the cell 2024A, the entire row content or partial content of the second data row associated with the cell 2024A can be displayed at the associated position (e.g., above or below) of the cell 2024A (when it exceeds the page display range, partial content can be displayed and scrolled through the scroll bar). Another example is that when clicking or double-clicking the cell 2024A, it can jump to the page of the second data table 212 to display the second data table 212 and highlight the entire row content or partial content of the second data row associated with the cell 2024A therein.

[0211] In some embodiments, when the above-mentioned association operation is a two-way association, in addition to realizing the functions that can be realized by the above-mentioned one-way association, the information of the target data row 2022A can also be viewed in the second data row corresponding to the target data row 2022A in the second data table 212, so that it is convenient for the user to view the detailed information of the target data row 2022A in other data tables that are two-way associated with the target data row 2022A. Taking the first data table 202 as an order table and the second data table 212 as a commodity table as an example, refer to Figure 2B As shown, the second data row 2122 of the second data table 212 may include an order number field. When the target data row 2022A is two-way associated with this second data row, the order number field of the second data row 2122 will be correspondingly associated with the target data row 2022A of the order table (the first data table 202). Thus, the user can view the information of the target data row 2022A associated therewith by triggering the cell corresponding to the order number field of this second data row. Similarly, the entire content or partial content of the target data row 2022A can be viewed through a hover operation, and by clicking or double-clicking, it is possible to jump to the page of the first data table 202 to display the first data table 202 and highlight the entire content or partial content of the target data row 2022A associated with this cell therein.

[0212] It can be seen from the above embodiments that the embodiments of the present disclosure provide a method for processing tabular data. By integrating at least two recall strategies, the coverage rate and accuracy of data recommendation can be improved to a certain extent. In some embodiments, four complementary recall strategies are designed, and potential associated records (data rows) can be mined from different dimensions. In some embodiments, the union strategy is adopted to merge the recall results of each path, and through deduplication processing, a more comprehensive candidate set is obtained. In some embodiments, a multi-level sorting strategy is adopted, and according to the priority of the recall strategy and the sorting rules within the strategy, the candidate set is sorted to improve the accuracy of the recommendation result. In some embodiments, a start condition is set for each recall strategy to ensure that the corresponding strategy is started in a suitable scenario, improving the recommendation efficiency and accuracy. In some embodiments, in the case where all recall strategies are not triggered or no data is recalled after being triggered, a fallback strategy is adopted to ensure the availability of the recommendation system.

[0213] Some embodiments of the present disclosure can design and apply a variety of complementary recall strategies according to different data characteristics and application scenarios, and fuse the recall results of each path through a merging strategy, and finally present the optimal candidate set of associated records for the user according to the sorting strategy.

[0214] Some embodiments of the present disclosure can effectively improve the recall rate of associated records and avoid missing valuable information by integrating multiple recall strategies to mine potential associated records from different dimensions. Some embodiments of the present disclosure can effectively filter out noisy data and improve the accuracy and relevance of recommendation results by using multiple recall strategies to complement each other and combining LLM screening and multi-level sorting strategies. Some embodiments of the present disclosure can adaptively select and combine appropriate recall strategies according to different data characteristics and application scenarios to enhance the robustness of the system.

[0215] It can be understood that the above data recommendation algorithm can be applied not only to the scenarios described in the above embodiments. In the era of information explosion, when users are faced with a vast amount of data, it is crucial to quickly and accurately find the required information. Therefore, the above data recommendation algorithm can also be widely applied to various application scenarios, such as knowledge base construction, data analysis, intelligent assistants, etc.

[0216] It should be noted that the method of the embodiments of the present disclosure can be executed by a single device, such as a computer or a server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In such a distributed scenario, one of the multiple devices can only execute one or more steps of the method of the embodiments of the present disclosure, and these multiple devices will interact with each other to complete the described method.

[0217] It should be noted that some embodiments of the present disclosure have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the above embodiments and still achieve the desired results. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0218] Embodiments of the present disclosure also provide a computer device for implementing the above method 300. Figure 5 The hardware structure diagram of an exemplary computer device 500 provided by the embodiments of the present disclosure is shown. The computer device 500 can be used to implement Figure 1 the server 106, and can also be used to implement Figure 1 the terminal devices 102A and 102B. In some scenarios, this computer device 500 can also be used to implement Figure 1 the database server 108.

[0219] As Figure 5As shown, the computer device 500 may include: a processor 502, a memory 504, a network interface 506, a peripheral interface 508, and a bus 510. Among them, the processor 502, the memory 504, the network interface 506, and the peripheral interface 508 are communicatively connected to each other inside the computer device 500 via the bus 510.

[0220] The processor 502 may be a central processing unit (CPU), a graphics processing unit, a neural network processing unit (NPU), a microcontroller (MCU), a programmable logic device, a digital signal processor (DSP), an application specific integrated circuit (ASIC), or one or more integrated circuits. The processor 502 may be used to execute functions related to the technologies described in this disclosure. In some embodiments, the processor 502 may further include multiple processors integrated as a single logic component. For example, as Figure 5 shown, the processor 502 may include multiple processors 502a, 502b, and 502c.

[0221] The memory 504 may be configured to store data (e.g., instructions, computer code, etc.). As Figure 5 shown, the data stored in the memory 504 may include program instructions (e.g., program instructions for implementing the method 300 of the embodiments of this disclosure) and data to be processed (e.g., the memory may store configuration files of other modules, etc.). The processor 502 may also access the program instructions and data stored in the memory 504 and execute the program instructions to operate on the data to be processed. The memory 504 may include a volatile storage device or a non-volatile storage device. In some embodiments, the memory 504 may include a random access memory (RAM), a read-only memory (ROM), an optical disc, a magnetic disk, a hard disk, a solid state drive (SSD), a flash memory, a memory stick, etc.

[0222] The network interface 506 may be configured to provide communication with other external devices to the computer device 500 via a network. The network may be any wired or wireless network capable of transmitting and receiving data. For example, the network may be a wired network, a local wireless network (e.g., Bluetooth, WiFi, near field communication (NFC), etc.), a cellular network, the Internet, or a combination of the above. It can be understood that the type of the network is not limited to the above specific examples.

[0223] The peripheral interface 508 can be configured to connect the computer device 500 to one or more peripheral devices to enable information input and output. For example, the peripheral devices may include input devices such as keyboards, mice, touch pads, touch screens, microphones, various sensors, etc., and output devices such as displays, speakers, vibrators, indicator lights, etc.

[0224] The bus 510 can be configured to transfer information between various components of the computer device 500 (such as the processor 502, the memory 504, the network interface 506, and the peripheral interface 508), such as internal buses (e.g., processor - memory bus), external buses (USB ports, PCI - E bus), etc.

[0225] It should be noted that although the architecture of the computer device 500 shown above only shows the processor 502, the memory 504, the network interface 506, the peripheral interface 508, and the bus 510, in the specific implementation process, the architecture of the computer device 500 may further include other components necessary for normal operation. In addition, those skilled in the art can understand that the architecture of the computer device 500 above may also only include the components necessary to implement the solution of the embodiments of the present disclosure, and does not necessarily include all the components shown in the figure.

[0226] The embodiments of the present disclosure also provide a tabular data processing device. Figure 6 The schematic diagram of an exemplary device 600 provided by the embodiments of the present disclosure is shown. As Figure 6 shown, the device 600 can be used to implement the method 300 and may further include the following modules.

[0227] The first determination module 602 is configured to: in response to an associated data recommendation request for a target data row in the first data table, determine the characteristic information of the target data row;

[0228] The acquisition module 604 is configured to: according to the characteristic information, acquire at least one first data row that matches the characteristic information in the first data table and / or acquire at least one second data row that matches the characteristic information in the second data table associated with the first data table;

[0229] The second determination module 606 is configured to: according to the at least one first data row and / or the at least one second data row that match the characteristic information, determine the recommended associated data of the target data row;

[0230] The generation module 608 is configured to: according to the recommended associated data, generate an associated data recommendation result for the target data row.

[0231] In some embodiments, the first determination module 602 is configured to: extract a plurality of keywords from the target data row, and determine the feature information according to the plurality of keywords.

[0232] In some embodiments, the first determination module 602 is configured to: generate a feature vector corresponding to the target data row according to the plurality of keywords; and determine the feature information according to the plurality of keywords and the feature vector.

[0233] In some embodiments, the acquisition module 604 is configured to:

[0234] acquire at least one first candidate data row that matches the plurality of keywords in the first data table according to the plurality of keywords;

[0235] acquire at least one second candidate data row that matches the feature vector in the first data table according to the feature vector;

[0236] determine at least one first data row that matches the feature information according to the at least one first candidate data row and the at least one second candidate data row.

[0237] In some embodiments, the acquisition module 604 is configured to:

[0238] acquire at least one third candidate data row that matches the plurality of keywords in the second data table according to the plurality of keywords;

[0239] acquire at least one fourth candidate data row that matches the feature vector in the second data table according to the feature vector;

[0240] determine at least one second data row that matches the feature information according to the at least one third candidate data row and the at least one fourth candidate data row.

[0241] In some embodiments, the associated data recommendation request includes a request for recommending associated data for a target field of the target data row;

[0242] The second determination module 606 is configured to:

[0243] determine at least one data row in the second data table associated with the target field of at least one first data row that matches the feature information as at least one first candidate data row;

[0244] determine at least one second data row that matches the feature information as at least one second candidate data row;

[0245] Determine the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row and the at least one second candidate data row.

[0246] In some embodiments, the obtaining module 604 is configured to: determine at least one first edited data row generated within a first preset time period from the current time in the first data table; determine at least one data row in the second data table associated with the target field of the at least one first edited data row as at least one third candidate data row;

[0247] The second determination module 606 is configured to: determine the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, and the at least one third candidate data row.

[0248] In some embodiments, the obtaining module 604 is configured to:

[0249] In response to a third data row being already associated in the target field of the target data row, determine at least one other data row in the first data table that associates the third data row in the target field;

[0250] In response to the number of the at least one other data row being greater than a first number, determine at least one fourth data row that the at least one other data row also associates in the target field;

[0251] For each of the at least one fourth data row, determine whether the number of the at least one other data row associated with the fourth data row is greater than a second number;

[0252] Determine at least one of the at least one fourth data row with the number of the associated at least one other data row being greater than the second number as at least one fourth candidate data row;

[0253] The second determination module 606 is configured to: determine the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, the at least one third candidate data row, and the at least one fourth candidate data row.

[0254] In some embodiments, the second determination module 606 is configured to:

[0255] Input the at least one first candidate data row into a first screening model to obtain at least one first screened data row;

[0256] Input the at least one second candidate data row into a second screening model to obtain at least one second screened data row;

[0257] Determine the recommended associated data corresponding to the target field of the target data row according to the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row;

[0258] Wherein, both the first filtering model and the second filtering model are large language models. The first filtering model filters the at least one first candidate data row based on a first prompt, and the second filtering model filters the at least one second candidate data row based on a second prompt. The first prompt includes task information for filtering the at least one first candidate data row based on the association relationship between the first candidate data row and the first data table, and the second prompt includes task information for filtering the at least one second candidate data row based on the association relationship between the first data table and the second data table.

[0259] In some embodiments, the second determination module 606 is configured to:

[0260] Determine at least one duplicate data row among the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row;

[0261] Retain one of the at least one duplicate data row and delete the remaining duplicate data rows;

[0262] Determine the recommended associated data corresponding to the target field of the target data row according to the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row after deleting the remaining duplicate data rows.

[0263] In some embodiments, the second determination module 606 is configured to:

[0264] Sort the at least one first filtered data row according to the occurrence times of the same data rows in the at least one first filtered data row and / or the editing time of each first filtered data row, to obtain a first internal sorting corresponding to the at least one first filtered data row;

[0265] Sort the at least one second filtered data row according to the matching degree with the multiple keywords and / or the similarity with the feature vector corresponding to the at least one second filtered data row, to obtain a second internal sorting corresponding to the at least one second filtered data row;

[0266] Sort the at least one third candidate data row according to the occurrence times of the same data rows in the at least one third candidate data row and / or the editing time of each third candidate data row, to obtain a third internal sorting corresponding to the at least one third candidate data row;

[0267] Sort the at least one fourth candidate data row according to the occurrence probability of each fourth candidate data row in the at least one fourth candidate data row, to obtain a fourth internal sorting corresponding to the at least one fourth candidate data row;

[0268] Determine an external sorting among the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row;

[0269] Based on the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, sort the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row, to obtain the recommended associated data.

[0270] In some embodiments, the external sorting is in the order of the fourth internal sorting, the second internal sorting, the first internal sorting, and the third internal sorting in sequence;

[0271] The generating module 608 is configured to: based on the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, select a third quantity of data rows from the recommended associated data as the associated data recommendation result of the target data row.

[0272] In some embodiments, the obtaining module 604 is configured to: in response to the target data row not being empty and the first data table including the first data row that has performed an association operation, obtain the at least one first data row that matches the feature information in the first data table according to the feature information.

[0273] In some embodiments, the obtaining module 604 is configured to: in response to the target data row not being empty and the number of second data rows in the second data table being greater than a fourth quantity, obtain the at least one second data row that matches the feature information in the second data table associated with the first data table according to the feature information.

[0274] In some embodiments, the obtaining module 604 is configured to: in response to the at least one first edited data row generated within the first preset time period from the current time being included in the first data table and the association operation being performed on the target field of the at least one first edited data row, obtain the at least one first edited data row.

[0275] In some embodiments, the obtaining module 604 is configured to: in response to determining that the target field allows multiple second data rows to be associated, the proportion of the first data rows in the first data table whose target field has associated multiple second data rows being higher than a preset ratio, and the number of the first data rows in the first data table being greater than a fifth number, when a third data row has been associated in the target field of the target data row, determine at least one other data row in the first data table that has associated the third data row in the target field.

[0276] In some embodiments, the obtaining module 604 is configured to: in response to failing to obtain the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row, determine at least one second edited data row within the second preset time period from the current time in the second data table or the sixth number of third edited data rows closest to the current time as the recommended associated data of the target data row.

[0277] For the convenience of description, when describing the above device, various modules are described separately according to their functions. Of course, when implementing the present disclosure, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0278] The device in the above embodiments is used to implement the corresponding method 300 in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated herein.

[0279] Based on the same inventive concept, corresponding to the method in any of the above embodiments, the present disclosure further provides a non-volatile computer-readable storage medium containing a computer program, which, when executed by one or more processors, causes the one or more processors to execute the method 300.

[0280] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage, or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0281] The computer program stored in the storage medium of the above embodiment is used to cause the one or more processors to execute the method 300 described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0282] Based on the same inventive concept, corresponding to the method 300 in any of the above embodiments, the present disclosure also provides a computer program product, which includes a computer program. In some embodiments, the computer program is executable by one or more processors to cause the processors to execute the method 300. Corresponding to the execution subjects of the respective steps in the method 300, the processors that execute the corresponding steps can belong to the corresponding execution subjects.

[0283] The computer program product of the above embodiment is used to cause the processor to execute the method 300 described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0284] Those of ordinary skill in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples; under the concept of the present disclosure, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the embodiments of the present disclosure as described above, which are not provided in detail for the sake of brevity.

[0285] Additionally, for simplicity of explanation and discussion, and so as not to make the embodiments of the present disclosure difficult to understand, well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Further, the devices may be shown in block diagram form in order to avoid making the embodiments of the present disclosure difficult to understand, and this also takes into account the fact that details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present disclosure are to be implemented (i.e., these details should be fully within the understanding of those skilled in the art). In cases where specific details (such as circuits) are set forth to describe exemplary embodiments of the present disclosure, it will be apparent to those skilled in the art that the embodiments of the present disclosure may be practiced without these specific details or with variations of these specific details. Accordingly, these descriptions should be regarded as illustrative rather than restrictive.

[0286] Although the present disclosure has been described in connection with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. For example, other memory architectures (such as dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0287] Embodiments of the present disclosure are intended to cover all such alternatives, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A method for processing tabular data, comprising: In response to an associated data recommendation request for a target data row in a first data table, determining feature information of the target data row; According to the feature information, obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in a second data table associated with the first data table; According to the at least one first data row and / or the at least one second data row that match the feature information, determining recommended associated data for the target data row; Generating an associated data recommendation result for the target data row according to the recommended associated data.

2. The method according to claim 1, wherein The determining the feature information of the target data row further includes: Extracting a plurality of keywords from the target data row and determining the feature information according to the plurality of keywords.

3. The method according to claim 2, wherein, The determining the feature information according to the plurality of keywords further includes: Generating a feature vector corresponding to the target data row according to the plurality of keywords; Determining the feature information according to the plurality of keywords and the feature vector.

4. The method according to claim 3, wherein The obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in a second data table associated with the first data table according to the feature information further includes: Obtaining at least one first candidate data row that matches the plurality of keywords in the first data table according to the plurality of keywords; Obtaining at least one second candidate data row that matches the feature vector in the first data table according to the feature vector; Determining the at least one first data row that matches the feature information according to the at least one first candidate data row and the at least one second candidate data row.

5. The method according to claim 3, wherein The obtaining at least one first data row that matches the feature information in the first data table and / or obtaining at least one second data row that matches the feature information in a second data table associated with the first data table according to the feature information further includes: Obtaining at least one third candidate data row that matches the plurality of keywords in the second data table according to the plurality of keywords; Obtaining at least one fourth candidate data row that matches the feature vector in the second data table according to the feature vector; Determining the at least one second data row that matches the feature information according to the at least one third candidate data row and the at least one fourth candidate data row.

6. The method according to claim 3, wherein, The associated data recommendation request includes a request for recommending associated data for a target field of the target data row; The determining the recommended associated data for the target data row according to the at least one first data row and / or the at least one second data row that match the feature information further includes: Determining at least one first candidate data row as at least one first alternative data row in the second data table associated with the target field of the at least one first data row that matches the feature information; Determine the at least one second data row that matches the feature information as at least one second candidate data row; Determine the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row and the at least one second candidate data row.

7. The method according to claim 6, further comprising: Determine at least one first edited data row generated within a first preset time period from the current time in the first data table; Determine at least one data row in the second data table associated with the target field of the at least one first edited data row as at least one third candidate data row; The determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row and the at least one second candidate data row further includes: determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, and the at least one third candidate data row.

8. The method according to claim 7, further comprising: In response to a third data row being associated with the target field of the target data row, determine at least one other data row in the first data table that associates the third data row in the target field; In response to the number of the at least one other data row being greater than a first number, determine at least one fourth data row that the at least one other data row also associates in the target field; For each of the at least one fourth data row, determine whether the number of the at least one other data row associated with the fourth data row is greater than a second number; Determine at least one of the at least one fourth data row whose number of the associated at least one other data row is greater than the second number as at least one fourth candidate data row; The determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, and the at least one third candidate data row further includes: determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, the at least one third candidate data row, and the at least one fourth candidate data row.

9. The method according to claim 8, wherein The determining the recommended associated data corresponding to the target field of the target data row according to the at least one first candidate data row, the at least one second candidate data row, the at least one third candidate data row, and the at least one fourth candidate data row further includes: Input the at least one first candidate data row into a first screening model to obtain at least one first screened data row; Input the at least one second candidate data row into a second screening model to obtain at least one second screened data row; Determine the recommended associated data corresponding to the target field of the target data row according to the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row; Among them, both the first filtering model and the second filtering model are large language models. The first filtering model filters the at least one first candidate data row based on a first prompt. The second filtering model filters the at least one second candidate data row based on a second prompt. The first prompt includes task information for filtering the at least one first candidate data row based on the association relationship between the first candidate data row and the first data table. The second prompt includes task information for filtering the at least one second candidate data row based on the association relationship between the first data table and the second data table.

10. The method according to claim 9, wherein, The determining the recommended associated data corresponding to the target field of the target data row according to the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row further includes: Determine at least one duplicate data row among the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row; Retain one of the at least one duplicate data row and delete the remaining duplicate data rows; Determine the recommended associated data corresponding to the target field of the target data row according to the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row after deleting the remaining duplicate data rows.

11. The method according to claim 9, wherein, The determining the recommended associated data corresponding to the target field of the target data row according to the at least one first filtered data row, the at least one second filtered data row, the at least one third candidate data row, and the at least one fourth candidate data row further includes: Sort the at least one first filtered data row according to the occurrence times of the same data rows in the at least one first filtered data row and / or the editing time of each first filtered data row to obtain a first internal sorting corresponding to the at least one first filtered data row; Sort the at least one second filtered data row according to the matching degree with the multiple keywords and / or the similarity with the feature vector corresponding to the at least one second filtered data row to obtain a second internal sorting corresponding to the at least one second filtered data row; Sort the at least one third candidate data row according to the occurrence times of the same data rows in the at least one third candidate data row and / or the editing time of each third candidate data row to obtain a third internal sorting corresponding to the at least one third candidate data row; Sort the at least one fourth candidate data row according to the occurrence probability of each fourth candidate data row in the at least one fourth candidate data row, to obtain a fourth internal sorting corresponding to the at least one fourth candidate data row; Determine an external sorting among the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row; Based on the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, sort the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row, to obtain the recommended associated data.

12. The method according to claim 11, wherein, The external sorting is in the order of the fourth internal sorting, the second internal sorting, the first internal sorting, and the third internal sorting in sequence; The generating the associated data recommendation result of the target data row according to the recommended associated data further includes: based on the external sorting, the first internal sorting, the second internal sorting, the third internal sorting, and the fourth internal sorting, select a third number of data rows from the recommended associated data as the associated data recommendation result of the target data row.

13. The method according to claim 1, wherein, Obtaining at least one first data row that matches the feature information in the first data table according to the feature information and / or obtaining at least one second data row that matches the feature information in a second data table associated with the first data table further includes: In response to the target data row not being empty and the first data table including the first data row for which an association operation has been performed, obtaining the at least one first data row that matches the feature information in the first data table according to the feature information; and / or In response to the target data row not being empty and the number of the second data rows in the second data table being greater than a fourth number, obtaining the at least one second data row that matches the feature information in the second data table associated with the first data table according to the feature information.

14. The method according to claim 7, wherein, The determining at least one first edited data row generated within a first preset time period from the current time in the first data table further includes: In response to the first data table including the at least one first edited data row generated within the first preset time period from the current time and the target field of the first edited data row having had an association operation performed, obtaining the at least one first edited data row.

15. The method according to claim 8, wherein, The responding to a third data row being associated in the target field of the target data row and determining at least one other data row in the first data table that associates the third data row in the target field further includes: In response to determining that the target field allows association of multiple second data rows, the proportion of the first data rows in the first data table that have associated multiple second data rows in the target field is higher than a preset ratio, and the number of the first data rows in the first data table is greater than a fifth number, when a third data row has been associated in the target field of the target data row, determine at least one other data row in the first data table that associates the third data row in the target field.

16. The method according to claim 9, further comprising: In response to failure to obtain the at least one first screened data row, the at least one second screened data row, the at least one third candidate data row, and the at least one fourth candidate data row, determine at least one second edited data row within a second preset time period from the current time or the sixth number of third edited data rows closest to the current time in the second data table as the recommended associated data of the target data row.

17. A table data processing device, comprising: A first determination module configured to: in response to an associated data recommendation request for a target data row in a first data table, determine the characteristic information of the target data row; An acquisition module configured to: according to the characteristic information, acquire at least one first data row that matches the characteristic information in the first data table and / or acquire at least one second data row that matches the characteristic information in a second data table associated with the first data table; A second determination module configured to: according to the at least one first data row and / or the at least one second data row that match the characteristic information, determine the recommended associated data of the target data row; A generation module configured to: according to the recommended associated data, generate an associated data recommendation result for the target data row.

18. A computer device, comprising one or more processors, a memory; and one or more programs, wherein the one or more programs are stored in the memory and executed by the one or more processors, and the one or more programs include instructions for executing the method according to any one of claims 1 to 16.

19. A non-volatile computer-readable storage medium containing a computer program, when the computer program is executed by one or more processors, causes the one or more processors to execute the method according to any one of claims 1 to 16.

20. A computer program product, comprising one or more computer programs, when the one or more computer programs are executed by one or more processors, implement the steps of the method according to any one of claims 1 to 16.

Images