A standard adaptive execution method and device based on data standard management

Abstract

The application provides a standard adaptive execution method and device based on data standard management, and belongs to the technical field of data management. The method comprises the following steps: in response to a standardization request of a user to a business system, a corresponding mapping relationship library is generated based on a preset root library and a source system field. The mapping relationship library at least comprises a mapping relationship between the source system field and a standard field established based on a preset semantic matching model. Each source data value of the business system is converted according to the mapping relationship library, and before being written into a target library, real-time verification is performed based on a preset multi-level value domain enumeration constraint to intercept conversion abnormal data and generate an abnormal snapshot according to the verification result. When it is determined that there is a data standard attribute change event or there is an abnormal snapshot that meets a preset condition, a corresponding update strategy type is determined based on the change content of the data standard version change or the root cause classification result of each abnormal snapshot, so that the standardization execution is performed according to the corresponding update strategy.

Description

Technical Field

[0001] This application relates to the field of data management technology, and in particular to a standard adaptive execution method and device based on data standard management. Background Technology

[0002] With the continuous development of digitalization, in the current field of data governance, especially for data-intensive enterprises (such as finance and healthcare), there is a common challenge of unifying the data from scattered business systems to enterprise-level data standards.

[0003] Currently, existing solutions mainly rely on manual configuration of field mapping relationships and use ETL tools to complete the source data transformation and writing of the standard library. This approach requires a lot of professional human resources, and due to the difference in understanding between business personnel and technical personnel, the mapping error rate is high and the mapping configuration efficiency is low.

[0004] Furthermore, when the enterprise data standard version changes, the existing mapping chain needs to be completely reconstructed. Currently, the method of batch compliance verification using SQL scripts also results in untimely monitoring of non-compliant data, impacting the user experience. Summary of the Invention

[0005] This application provides a standard adaptive execution method and device based on data standard management, which solves the technical problems of current data standard management, such as excessive reliance on manual processes for labeling and mapping, low efficiency, inability to cope with the difficulties brought about by standard changes, and untimely monitoring of labeling quality, resulting in a poor user experience.

[0006] On one hand, embodiments of this application provide a standard adaptive execution method based on data standard management, the method being applied to a data standard management system; the method includes:

[0007] In response to user requests for standardization of the business system, a corresponding mapping relationship library is generated based on a preset root word library and source system fields; wherein, the mapping relationship library includes at least the mapping relationship between the source system fields and standard fields established based on a preset semantic matching model;

[0008] The source data values ​​of the business system are transformed according to the mapping relationship library, and real-time verification is performed based on the preset multi-level value domain enumeration constraints before being written to the target library, so as to intercept abnormal data and generate abnormal snapshots according to the verification results; the preset multi-level value domain enumeration constraints correspond to at least the following value domains: basic value domain, business value domain and standard value domain;

[0009] When it is determined that there is a data standard attribute change event or an abnormal snapshot that meets preset conditions, based on the change content of the data standard version change or the root cause classification result of each abnormal snapshot, the corresponding update strategy type is determined so as to perform standardized execution according to the corresponding update strategy; wherein, the update strategy type includes at least: field value conversion script and mapping rule update instruction.

[0010] In one implementation of this application, a corresponding mapping relationship library is generated based on a preset root word library and source system fields, specifically including:

[0011] The source system field is parsed according to the preset root word library to split the source system field into a root word sequence, and then input into the preset semantic matching model to construct the corresponding root word semantic vector; wherein, the root word semantic vector is obtained by concatenating entity mapping sub-vectors and relation mapping sub-vectors;

[0012] Calculate the vector similarity between the semantic vector of the word root and the feature vector of the preset standard field;

[0013] If the vector similarity is greater than a preset similarity threshold, a mapping relationship between the source system field and the corresponding standard field is established to generate the mapping relationship library.

[0014] In one implementation of this application, real-time verification is performed based on preset multi-level value range enumeration constraints to intercept and convert abnormal data and generate an abnormal snapshot according to the verification results, specifically including:

[0015] Based on the preset multi-level value range enumeration constraints, the source data values ​​are subjected to hierarchical verification.

[0016] When the source data value fails at any verification level, the writing of the source data value to the target database is terminated, and the abnormal snapshot is generated.

[0017] In one implementation of this application, hierarchical validation of the source data values ​​specifically includes:

[0018] First level: Determine whether the source data value conforms to the basic value range constraints, which include data type constraints, length constraints, and format constraints;

[0019] Second level: If the first level of verification is passed, it is determined whether the source data value is within the first enumerated value range allowed by the business value domain constraint;

[0020] Third level: If the second level of verification is passed, it is determined whether the source data value is within the second enumerated value range allowed by the standard value range constraint.

[0021] In one implementation of this application, the abnormal snapshot includes at least one or more of the following information: abnormal data value and its original metadata, specific value domain constraints violated, associated mapping relationship labels, and abnormal verification level identifiers.

[0022] In one implementation of this application, the method further includes:

[0023] When it is determined that the source data value does not meet the type matching condition corresponding to the data type constraint, a preset conversion rule is executed to convert the source data type corresponding to the source data value into the target field type; the preset conversion rule includes at least one or more of the following: string to numeric conversion, numeric to string conversion, date format standardization conversion, and boolean to integer mapping conversion.

[0024] In one implementation of this application, before determining that a data standard attribute change event exists, the method further includes:

[0025] Monitor the standard field template and determine whether the attributes of the standard field template have been modified;

[0026] If so, mark the standard set associated with the standard field model and generate the data standard attribute change event.

[0027] In one implementation of this application, determining the existence of an abnormal snapshot that meets preset conditions specifically includes:

[0028] The number of abnormal snapshots is accumulated, and the business standard type corresponding to each abnormal snapshot is determined;

[0029] If the number of abnormal snapshots is greater than a predetermined value and / or if there are abnormal snapshots of the selected business standard type, the preset condition is determined to be met.

[0030] In one implementation of this application, based on the changes in the data standard version or the root cause classification results of each of the abnormal snapshots, a corresponding update strategy type is determined, and standardized execution is performed according to the corresponding update strategy, specifically including:

[0031] Based on the changed content or the root cause classification result, the update strategy type is determined through strategy determination rules; wherein, the strategy determination rules include: when the changed content involves a value range or data structure, the update strategy type is determined to be a field value conversion script; when the changed content only involves semantic rules or associations, the update strategy type is determined to be a mapping rule update instruction; when the root cause classification result is a standard definition defect, the update strategy type is determined to be a field value conversion script; when the root cause classification result is a mapping rule error, the update strategy type is determined to be a mapping rule update instruction.

[0032] When the update strategy type is field value conversion script, generate and execute the field value conversion script to convert the source data value;

[0033] When the update strategy type is the mapping rule update instruction, the mapping rule update instruction is generated and the parameters of the preset semantic matching model are adjusted.

[0034] On the other hand, embodiments of this application also provide a standard adaptive execution device based on data standard management, the device being applied to a data standard management system; the device includes:

[0035] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a standard adaptive execution method based on data standard management as described above.

[0036] Compared with the prior art, the significant advantages of this application are as follows:

[0037] The above technical solution automatically generates a mapping relationship library through a preset semantic matching model, replacing manual configuration of field mappings. It also utilizes a multi-level value range verification mechanism to intercept anomalies, significantly improving the efficiency and accuracy of data standardization. Furthermore, it determines update strategies based on the root causes of data standard attribute change events or abnormal snapshots, achieving adaptive anomaly management. This solves the current technical problems in data standard management, such as excessive reliance on manual labeling and mapping, low efficiency, inability to cope with standard changes, and untimely monitoring of labeling quality, leading to a poor user experience. Attached Figure Description

[0038] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0039] Figure 1 This is a flowchart illustrating a standard adaptive execution method based on data standard management in an embodiment of this application;

[0040] Figure 2 This is a schematic diagram of the structure of a standard adaptive execution device based on data standard management in an embodiment of this application. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0042] This application provides a standard adaptive execution method and device based on data standard management to solve the technical problems of current data standard management, which relies too much on manual labor for labeling and mapping, resulting in low efficiency, inability to cope with the problems brought about by standard changes, and untimely monitoring of labeling quality, leading to a poor user experience.

[0043] The various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0044] This application provides a standard adaptive execution method based on data standard management, such as... Figure 1 As shown, this method is applied to a data standard management system, and the method may include steps S101-S103:

[0045] S101, the server responds to the standardized request from the user to the business system and generates the corresponding mapping relationship library based on the preset root word library and source system fields.

[0046] The mapping relationship library includes at least the mapping relationship between source system fields and standard fields established based on a preset semantic matching model.

[0047] It should be noted that the server, as the execution subject of the standard adaptive execution method based on data standard management, exists only as an example, and the execution subject is not limited to the server. This application does not make any specific limitation in this regard.

[0048] The data standard management system includes at least the following modules: a code table management module for managing multi-level code table directories and code values, supporting multi-level code table directory management, allowing users to create up to 5 levels of code table directories, with up to 10,000 code values ​​added to each directory; a root word management module for managing business terms, physical tables, field naming conventions, and their standardized translations; a standard template management module for defining the attribute information required when creating data standards; a standard set management module for managing a set of similar or consistent standards related to business meaning; a data standard management module for the standardized definition and application of data standards; and a benchmark evaluation management module for defining the matching relationship between standard attributes and asset object metadata fields, and evaluating the mapping relationship.

[0049] Users can establish communication connections with the server through user devices such as mobile phones and computers, and send standardization requests based on the standardization of business systems. For example, standardizing data from a medical system or data from a financial system.

[0050] In this embodiment of the application, a corresponding mapping relationship library is generated based on a preset root word library and source system fields, specifically including:

[0051] The source system fields are parsed based on a pre-defined root word library to split them into root word sequences, which are then input into a pre-defined semantic matching model to construct corresponding root word semantic vectors. These root word semantic vectors are obtained by concatenating entity mapping sub-vectors and relation mapping sub-vectors. The vector similarity between the root word semantic vectors and the feature vectors of pre-defined standard fields is calculated. If the vector similarity is greater than a pre-defined similarity threshold, a mapping relationship between the source system fields and the corresponding standard fields is established to generate a mapping relationship library.

[0052] In other words, this application can parse source system fields using a preset root word library. This can be understood as splitting the source system fields into one or more root words that are consistent with the root words existing in the preset root word library, and assembling them into a root word sequence. The preset root word library can be obtained through the aforementioned root word management module.

[0053] Subsequently, entity mapping sub-vectors are constructed based on entity mapping rules using a pre-defined semantic matching model, and relation mapping sub-vectors are constructed based on relation mapping rules. The entity mapping sub-vectors and relation mapping sub-vectors are then weighted and concatenated according to pre-defined concatenation weights to obtain the root semantic vector.

[0054] The preset semantic matching model can be a pre-trained deep learning model or other models; this application does not impose any specific limitations on it. The entity mapping rule assigns a preset entity weight to each root word in the field; the relation mapping rule calculates the association strength of the root words, forming a relation mapping sub-vector containing the association strength. The association strength can be obtained based on the product of the root word co-occurrence frequency and the preset business weight. The preset entity weight, preset business weight, and preset concatenation weight can all be obtained from expert experience; this application does not impose any specific limitations on them. Generating root semantic vectors that include entity mapping and relation mapping can effectively improve the matching accuracy of complex fields.

[0055] Subsequently, the server will calculate vector similarity using a preset semantic matching model. This can be done using the cosine similarity formula, or other similarity calculation formulas; no specific limitation is made here. When the vector similarity exceeds a preset similarity threshold, the server will select the standard field corresponding to that vector similarity as a mappable standard field for the source system field and establish a mapping relationship, thereby constructing a mapping relationship library containing these relationships. In other words, the semantic matching model uses entity mapping rules and relation mapping rules to construct the field feature vectors (word root semantic vectors) corresponding to the source system fields and calculates vector similarity to determine the mapping matching degree.

[0056] For example, the source system fields are: medical patient, clinical number, social insurance, and the corresponding root sequence is [patient, number, social insurance]. After performing the above mapping, the standard field is obtained as patient-social insurance, and finally the mapping relationship of (patient A's number is B, patient A has social insurance)-(patient-social insurance) is obtained.

[0057] If no source system field with a similarity threshold greater than the preset threshold is found, the server can generate a prompt message and send it to the user's terminal, allowing for manual intervention to construct the mapping relationship. The preset similarity threshold can be set by the user based on their actual usage scenario and is not specifically limited here.

[0058] S102, the server transforms the source data values ​​of the business system according to the mapping relationship library, and performs real-time verification based on the preset multi-level value domain enumeration constraints before writing to the target library, so as to intercept abnormal data and generate abnormal snapshots according to the verification results.

[0059] The preset multi-level value range enumeration constraint corresponds to at least the following value ranges: basic value range, business value range, and standard value range.

[0060] After generating the mapping database, the source data values ​​in the business system can be converted into standard data. However, during the mapping conversion process, issues such as incorrect or invalid mappings may occur, necessitating verification. In this embodiment, real-time verification is performed based on preset multi-level value range enumeration constraints to intercept abnormal conversion data and generate an abnormal snapshot based on the verification results. Specifically, this includes:

[0061] Based on preset multi-level value range enumeration constraints, the source data values ​​are subjected to hierarchical validation. When a source data value fails at any validation level, the writing of the source data value to the target database is terminated, and an abnormal snapshot is generated.

[0062] In other words, this application uses the rules corresponding to the pre-set multi-level value range enumeration constraints to perform multi-level verification on the source data values ​​of the conversion and intercept abnormal data.

[0063] This includes performing hierarchical validation on the source data values, specifically:

[0064] The first level: Determine whether the source data value conforms to the basic value range constraints, which include data type constraints, length constraints, and format constraints. That is, verify whether the data type, length, and format of the source data value conform to the rules. Data types can include integer, character, and date types; length constraints are in bytes; format constraints can be defined by regular expressions.

[0065] The second level: If the first level of verification passes, it determines whether the source data value is within the first enumerated value range allowed by the business value domain constraints. That is, if the basic value domain verification passes, it verifies whether the data value is within the first enumerated value range allowed by the source system's business rules. The first enumerated value range can be obtained through dynamic connection to the source system's metadata database and updated in real time, or it can be determined by the user based on expert experience and equipment; no specific limitations are specified here.

[0066] The third level: If the second-level validation passes, it determines whether the source data value is within the second enumerated value range allowed by the standard value range constraint. That is, if the business value range validation passes, it verifies whether the data value conforms to the second enumerated value range constraint defined by industry standards or enterprise standards. The second enumerated value range can be obtained by the user through synchronization with external industry standard libraries via a standard application programming interface (API), and this application does not specifically limit it.

[0067] When a data value fails at any validation level, writing to the target database is immediately terminated and an anomaly snapshot is generated. The anomaly snapshot includes at least one or more of the following information: the anomaly data value and its original metadata, the specific value domain constraint violated, the associated mapping relationship label (ID), and the anomaly validation level identifier.

[0068] For example, abnormal snapshot = {

[0069] "Abnormal data value": The original value, / / e.g., ID number '11010119990101ABCD'

[0070] "Specific value range constraint violated": "Basic value range / length constraint (18 bits required)"

[0071] "Mapping ID": "MAP_2024_ID_CARD", / / Associated with a specific mapping rule

[0072] "Abnormal Level": "Basic Value Range" / / Identifies the failure link

[0073] }

[0074] The mapping relationship ID is associated with the mapping relationship in the mapping relationship database mentioned above.

[0075] This application employs multi-level value range sequential verification, which not only verifies a single value range, improving verification accuracy, but also performs verification before writing to the target database, avoiding contamination of the target database after writing and affecting the subsequent implementation of data standards. Simultaneously, hierarchical filtering effectively reduces unnecessary computation.

[0076] Furthermore, for individual transformation errors that do not meet the corresponding conditions of the multi-level value range enumeration constraints, remedial rules can be set to avoid generating abnormal snapshots and causing standardized execution waits. Therefore, this application provides the following embodiments, including:

[0077] When it is determined that the source data value does not meet the type matching conditions corresponding to the data type constraint, a preset conversion rule is executed to convert the source data type corresponding to the source data value to the target field type. The preset conversion rule includes at least one or more of the following: string to numeric conversion, numeric to string conversion, date format standardization conversion, and boolean to integer mapping conversion.

[0078] In other words, when converting source data values, if the source data values ​​do not meet the type matching conditions in the data type constraints, such as differences between the source data value type and the target type in terms of character, integer, floating-point, date, or boolean types, resulting in failure to meet the category matching conditions, the server can call the preset conversion rules to convert the source data values ​​to the target field type, thus avoiding being intercepted as conversion error data.

[0079] For example, when the source data value (such as the string "123.45") does not match the target field data type (such as the floating-point type "DECIMAL"), the automatic format (string to numeric) conversion operation protocol is executed according to the preset conversion rules.

[0080] Among them, the conversion of strings to numeric types in this application removes non-numeric characters and converts them to numeric values, such as removing currency symbols ($￥€) and thousands separators (,), and converting negative strings to negative numeric values ​​(such as "(123)" to -123); the conversion of numeric types to strings: zero padding is performed according to the target length; the standardization conversion of date formats: the source date string is parsed according to the preset template and converted to the target format; the mapping conversion of boolean types to integer types: true values ​​are mapped to 1, and false values ​​are mapped to 0.

[0081] S103, when the server determines that there is a data standard attribute change event or an abnormal snapshot that meets the preset conditions, it determines the corresponding update strategy type based on the change content of the data standard version change or the root cause classification results of each abnormal snapshot, so as to perform standardized execution according to the corresponding update strategy.

[0082] The update strategy types include at least: field value conversion scripts and mapping rule update instructions.

[0083] In this embodiment of the application, before determining that a data standard attribute change event exists, the method further includes:

[0084] Monitor standard field templates and determine if their attributes have been modified. If a modification is confirmed, flag the standard set associated with the standard field model and generate a data standard attribute change event.

[0085] Specifically, monitoring standard field templates can be achieved through the data standard management module to determine whether the attributes in the standard field templates have changed. If there are changes, the standard set involved in that attribute can be marked, and a data standard attribute change event can be generated to update the strategy in a timely manner and avoid abnormal data conversion caused by subsequent template changes.

[0086] In this application embodiment, determining the existence of abnormal snapshots that meet preset conditions specifically includes:

[0087] The system accumulates the number of abnormal snapshots and determines the business standard type corresponding to each abnormal snapshot. If the number of abnormal snapshots exceeds a predetermined value and / or an abnormal snapshot of a selected business standard type exists, the system determines that a preset condition has been met.

[0088] In other words, the server will monitor the number of abnormal snapshots generated and / or the corresponding business standard types of these snapshots. These business standard types can be pre-set on the server by users or experts. Different abnormal snapshots can correspond to different business standard types, and users can select one or more business standard types. When abnormal snapshots of these selected business standard types appear, it is determined that an abnormal snapshot meeting the preset conditions exists. If the number of abnormal snapshots generated exceeds a predetermined value, the preset conditions are determined to be met. Alternatively, the server can combine both conditions; if the number of generated abnormal snapshots exceeds the predetermined value and an abnormal snapshot of a selected business standard type exists, the preset conditions are considered met. The specific rules for determining the preset conditions can be set by the user according to the actual usage scenario; this application does not impose specific limitations on this.

[0089] In this embodiment of the application, based on the changes in the data standard version or the root cause classification results of each abnormal snapshot, a corresponding update strategy type is determined, and standardized execution is performed according to the corresponding update strategy, specifically including:

[0090] Based on the changes or root cause classification results, the update strategy type is determined using strategy determination rules. These rules include: when the changes involve value ranges or data structures, the update strategy type is determined to be a field value transformation script; when the changes only involve semantic rules or relationships, the update strategy type is determined to be a mapping rule update instruction; when the root cause classification result is a standard definition defect, the update strategy type is determined to be a field value transformation script; and when the root cause classification result is a mapping rule error, the update strategy type is determined to be a mapping rule update instruction. When the update strategy type is a field value transformation script, the script is generated and executed to transform the source data values. When the update strategy type is a mapping rule update instruction, the instruction is generated and the parameters of the preset semantic matching model are adjusted.

[0091] In other words, the server can select the appropriate update strategy by changing the content and root cause classification results, thereby generating field value transformation scripts or updating the parameters of the preset semantic matching model.

[0092] Among these changes, the types of changes corresponding to the above-mentioned changes include: changes in value range, which can be determined by adding or removing enumerated values ​​or changing data types; changes in data results, which can be determined by splitting / merging fields or adding constraints; changes in semantic rules, which can be determined by renaming fields or updating the thesaurus; and changes in relationships, which can be determined by adding / deleting relationships between fields. The above root cause classification results can be obtained through clustering. For example, by using a representative density-based clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN) to perform cluster analysis on anomaly snapshots, standard definition defect classes can be obtained: the same standard field reports errors in multiple systems; and mapping rule error classes: specific source systems frequently violate business value ranges.

[0093] The above technical solution automatically generates a mapping relationship library through a preset semantic matching model, replacing manual configuration of field mappings. It also utilizes a multi-level value range verification mechanism to intercept anomalies, significantly improving the efficiency and accuracy of data standardization. Furthermore, it determines update strategies based on the root causes of data standard attribute change events or abnormal snapshots, achieving adaptive anomaly management. This solves the current technical problems in data standard management, such as excessive reliance on manual labeling and mapping, low efficiency, inability to cope with standard changes, and untimely monitoring of labeling quality, leading to a poor user experience.

[0094] Figure 2 A schematic diagram of a standard adaptive execution device based on data standard management is provided for an embodiment of this application, as shown below. Figure 2 As shown, the device is used in a data standards management system. The device includes:

[0095] At least one processor; and a memory communicatively connected to the at least one processor. The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to:

[0096] In response to user requests for standardization of the business system, a corresponding mapping relationship library is generated based on a preset root word library and source system fields. This mapping relationship library includes at least the mapping relationships between source system fields and standard fields established based on a preset semantic matching model. The source data values ​​of the business system are transformed according to the mapping relationship library. Before being written to the target library, real-time verification is performed based on preset multi-level value domain enumeration constraints to intercept abnormal data and generate abnormal snapshots based on the verification results. The preset multi-level value domain enumeration constraints correspond to at least the following value domains: basic value domain, business value domain, and standard value domain. When a data standard attribute change event is determined or an abnormal snapshot that meets preset conditions exists, an appropriate update strategy type is determined based on the change content of the data standard version change or the root cause classification results of each abnormal snapshot, and standardized execution is performed according to the corresponding update strategy. The update strategy type includes at least: field value conversion scripts and mapping rule update instructions.

[0097] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0098] The devices and methods provided in this application are one-to-one correspondences. Therefore, the devices also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices will not be repeated here.

[0099] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0100] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A standard adaptive execution method based on data standard management, characterized in that, The method is applied to a data standard management system; the method includes: In response to user requests for standardization of the business system, a corresponding mapping relationship library is generated based on a preset root word library and source system fields. This mapping relationship library includes at least the mapping relationship between the source system fields and standard fields established based on a preset semantic matching model. Specifically, this includes: parsing the source system fields according to the preset root word library to split them into root word sequences, and inputting these sequences into the preset semantic matching model to construct corresponding root word semantic vectors; wherein the root word semantic vectors are obtained by concatenating entity mapping sub-vectors and relation mapping sub-vectors; calculating the vector similarity between the root word semantic vectors and the feature vectors of the preset standard fields; if the vector similarity is greater than a preset similarity threshold, establishing a mapping relationship between the source system fields and the corresponding standard fields to generate the mapping relationship library; constructing entity mapping sub-vectors based on entity mapping rules using the preset semantic matching model, and constructing relation mapping sub-vectors based on relation mapping rules; the entity mapping rule assigns a preset entity weight to each root word in the field; the relation mapping rule calculates the association strength between root words, forming a relation mapping sub-vector containing the association strength, which is obtained based on the product of the root word co-occurrence frequency and a preset business weight. The source data values ​​of the business system are transformed according to the mapping relationship library, and real-time verification is performed based on the preset multi-level value domain enumeration constraints before being written to the target library. This is done to intercept abnormal data and generate an abnormal snapshot based on the verification results. The preset multi-level value domain enumeration constraints correspond to at least the following value domains: basic value domain, business value domain, and standard value domain. Specifically, this includes: performing hierarchical verification on the source data values ​​according to the preset multi-level value domain enumeration constraints: First level: determining whether the source data values ​​conform to the basic value domain constraints, which include data type constraints, length constraints, and format constraints; Second level: Level 1: If the first level of verification passes, it is determined whether the source data value is within the first enumerated value range allowed by the business value domain constraint; Level 2: If the second level of verification passes, it is determined whether the source data value is within the second enumerated value range allowed by the standard value domain constraint; When the source data value fails at any verification level, the writing of the source data value to the target database is terminated, and the abnormal snapshot is generated; The abnormal snapshot includes at least one or more of the following information: abnormal data value and its original metadata, the specific value domain constraint condition violated, the associated mapping relationship label, and the abnormal verification level identifier; When a data standard attribute change event is identified or an abnormal snapshot meeting preset conditions is identified, an appropriate update strategy type is determined based on the change content of the data standard version change or the root cause classification results of each abnormal snapshot, so as to perform standardized execution according to the corresponding update strategy. The update strategy type includes at least: a field value conversion script and a mapping rule update instruction. Specifically, it includes: determining the update strategy type based on the change content or the root cause classification results through strategy determination rules. The strategy determination rules include: determining the update strategy type as a field value conversion script when the change content involves a value range or data structure; determining the update strategy type as a mapping rule update instruction when the change content only involves semantic rules or relationships; determining the update strategy type as a field value conversion script when the root cause classification result is a standard definition defect; and determining the update strategy type as a mapping rule update instruction when the root cause classification result is a mapping rule error. When the update strategy type is a field value conversion script, the field value conversion script is generated and executed to convert the source data value. When the update strategy type is a mapping rule update instruction, the mapping rule update instruction is generated and the parameters of the preset semantic matching model are adjusted.

2. The standard adaptive execution method based on data standard management according to claim 1, characterized in that, The method further includes: When it is determined that the source data value does not meet the type matching condition corresponding to the data type constraint, a preset conversion rule is executed to convert the source data type corresponding to the source data value into the target field type; the preset conversion rule includes at least one or more of the following: string to numeric conversion, numeric to string conversion, date format standardization conversion, and boolean to integer mapping conversion.

3. The standard adaptive execution method based on data standard management according to claim 1, characterized in that, Before determining that a data standard attribute change event exists, the method further includes: Monitor the standard field template and determine whether the attributes of the standard field template have been modified; If so, mark the standard set associated with the standard field model and generate the data standard attribute change event.

4. The standard adaptive execution method based on data standard management according to claim 1, characterized in that, The existence of abnormal snapshots that meet preset conditions is confirmed, specifically including: The number of abnormal snapshots is accumulated, and the business standard type corresponding to each abnormal snapshot is determined; If the number of abnormal snapshots is greater than a predetermined value and / or if there are abnormal snapshots of the selected business standard type, the preset condition is determined to be met.

5. A standard adaptive execution device based on data standard management, characterized in that, The device is used in a data standards management system; the device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform a standard adaptive execution method based on data standard management as described in any one of claims 1-4.

Images