Data processing methods and related products for graph databases

Abstract

This invention provides a data processing method and related products for graph databases. The data processing method includes: acquiring a document query request; matching multiple candidate document entities in the graph database using graph retrieval based on the document query request; acquiring the document query vector of the document query request; acquiring the content attributes of each candidate document entity, where the content attributes are in vector form and used to represent the content of the corresponding candidate document entity; and using vector retrieval to match one or more of the content attributes of each candidate document entity with the highest similarity to the document query vector, which are then used as the target document entity. In this method, graph retrieval and vector retrieval support and cooperate with each other, achieving the goal of improving the overall accuracy and efficiency of document queries in graph databases.

Description

Technical Field

[0001] This invention relates to the field of database technology, and in particular to a data processing method for graph databases, a computer-readable storage medium, a computer program product, and a computer device. Background Technology

[0002] With the explosive growth of unstructured data, how to extract unstructured data into structured data for storage and build connections between knowledge has become a crucial issue. This has led to the concept of multimodal data, designed to represent various data types. Currently, there are two main solutions: the first is the vector database approach, which transforms unstructured data into embedded vectors using an embedding model and stores them in the vector database. Retrieval then searches for the closest approximate vector data in the vector database. The second is the graph database approach, which extracts unstructured data into entities and establishes relationships between these entities. Entities themselves possess attributes, and the graph model can concisely establish these relationships.

[0003] However, vector databases require significant computational resources for queries, while graph databases lack semantic capture during queries. Summary of the Invention

[0004] One objective of this invention is to provide a data processing method for graph databases, a computer-readable storage medium, a computer program product, and a computer device, enabling multi-modal retrieval of graph databases, comprehensively utilizing graph retrieval and vector retrieval, and improving the accuracy and efficiency of document queries in graph databases.

[0005] Specifically, according to one aspect of the present invention, the present invention provides a data processing method for a graph database, comprising: Get the document query request; Based on the document query request, graph retrieval is used to match multiple candidate document entities in the graph database; Obtain the document query vector of the document query request, wherein the document query vector is used to characterize the content of the expected document of the document query request; Obtain the content attributes of each candidate document entity, wherein the content attributes are in vector form and are used to represent the content of the corresponding candidate document entity; Using vector retrieval, one or more of the content attributes of each candidate document entity are matched to the document query vector with the highest similarity, and these are selected as the target document entities.

[0006] Optionally, the step of matching multiple candidate document entities in the graph database using graph retrieval based on the document query request includes: Obtain the graph index of the entity nodes in the graph database, where the entity nodes include document nodes and author nodes; Obtain the structured relationships between multiple entities in the document query request, where the entities include document entities and author entities; Based on the structured relationships and the graph index, a graph traversal operation is performed to match multiple candidate document entities in the graph database.

[0007] Optionally, before the step of obtaining the document query vector of the document query request, the method further includes: Construct summary attributes for multiple document nodes of the graph database. The summary attributes are descriptive text used to describe the summary of the document entity corresponding to the document node. Using an embedding model, the embedding vectors of each of the aforementioned summary attributes are obtained; Based on the embedding vectors, the content attributes of each document node are constructed.

[0008] Optionally, obtaining the document query vector of the document query request includes: Obtain the title of the desired document in the document query request; Match the document node in the graph database that corresponds to the title of the desired document; Obtain the content attributes of the document node corresponding to the desired document, and use them as the document query vector.

[0009] Optionally, obtaining the content attributes of each of the candidate document entities includes: The document nodes corresponding to each candidate document entity are obtained from the graph database, and their content attributes are read.

[0010] Optionally, after the step of constructing the content attributes of each of the document nodes, the method further includes: Construct a vector index of the content attributes of the document nodes in the graph database.

[0011] Optionally, the use of vector retrieval, which involves matching the multiple candidate document entities with the highest similarity to the document query vector among their content attributes, and selecting these as target document entities, includes: Using the vector index, calculate the similarity score between the content attribute of each candidate document entity and the document query vector; The calculated similarity scores are sorted to obtain the N highest similarity scores, where N is an integer greater than 1; The candidate document entities corresponding to the N highest similarity scores are taken as the target document entities.

[0012] According to another aspect of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the data processing method for the graph database described above.

[0013] According to another aspect of the present invention, a computer program product is also provided, comprising a computer program that, when executed by a processor, implements the steps of the data processing method for the graph database described above.

[0014] According to another aspect of the present invention, a computer device is also provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the data processing method for the graph database described above.

[0015] The graph database data processing method of this invention, by using graph retrieval, quickly, accurately, and efficiently matches multiple candidate document entities in the graph database, significantly narrowing the data range requiring vector retrieval. By additionally setting vector-form content attributes for each document node in the graph database, and using vector retrieval to perform semantic similarity searches on the content attributes of multiple candidate document entities, semantics are accurately captured with minimal resource overhead, resulting in one or more target document entities with the highest similarity. Graph retrieval and vector retrieval support and cooperate with each other, achieving the goal of improving the overall accuracy and efficiency of document queries in the graph database.

[0016] The above and other objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0017] The following sections will describe some specific embodiments of the invention in detail by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or portions. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a flowchart illustrating a data processing method according to an embodiment of the present invention; Figure 2 This is a schematic flowchart of a data processing method for graph retrieval according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating the construction of content attributes and vector indexes in a data processing method according to an embodiment of the present invention; Figure 4 This is a schematic flowchart of a data processing method for obtaining document query vectors according to an embodiment of the present invention; Figure 5 This is a schematic flowchart of a data processing method for vector retrieval according to an embodiment of the present invention; Figure 6 This is a flowchart illustrating the construction of content attributes in a data processing method according to another embodiment of the present invention; Figure 7 This is a schematic diagram of a computer program product according to an embodiment of the present invention; Figure 8 This is a schematic diagram of a computer-readable storage medium according to an embodiment of the present invention; and Figure 9 This is a schematic diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0018] The purpose of the graph database data processing method in this embodiment is to achieve multi-modal retrieval of graph databases, and to improve the accuracy and efficiency of document querying of graph databases by comprehensively using graph retrieval and vector retrieval.

[0019] Figure 1 This is a flowchart illustrating a data processing method for a graph database according to an embodiment of the present invention. The method generally includes: S100, retrieve document query request; S200: Based on the document query request, use graph retrieval to match multiple candidate document entities in the graph database; S300, Obtain the document query vector of the document query request. The document query vector is used to represent the content of the expected document in the document query request. S400, obtain the content attributes of each candidate document entity. The content attributes are in vector form and are used to represent the content of the corresponding candidate document entity. S500 uses vector retrieval to match one or more of the most similar entities to the document query vector in the content attributes of each candidate document entity, and uses them as the target document entity.

[0020] Graph databases are databases used for storing, querying, and updating graph-structured data. They are based on graph theory in mathematics and use graph models for data storage. Graph models are typically described using triples (nodes, relations, and properties). Nodes (also called entity nodes) represent entity information in the graph and can contain one or more properties. Nodes are connected using relations. Relations (also called edges) connect nodes and can have one or more properties; there can be multiple, even recursive, relations between nodes. Properties are typically named values, where the name (or key) is usually a string. Properties can be indexed and constrained, and composite indexes can be created from multiple properties.

[0021] In this embodiment, the graph database can be used to store, query, and update document data (e.g., paper data). Specifically, the graph database can set multiple document (Paper) nodes and author (Author) nodes. The document node is used to store document entities and can be set with multiple attributes, such as title, abstract, and publication year. The author node is used to store author entities and can be set with attributes such as name and affiliation.

[0022] Relationship edges are set between corresponding nodes, including writing relationships (represented by [:AUTHORED]) and mentoring relationships (represented by [:ADVISED]). The writing relationship connects the author node and the corresponding document node, while the mentoring relationship connects the tutor (author node) and their student (author node).

[0023] In this embodiment, the graph database also sets content attributes for each document node. These content attributes characterize the content of the corresponding candidate document entity and are stored as vectors. For example, the graph database can use the summary attribute of a document node to convert the summary content text into vector values. The graph database can also use manual or AI tools to summarize the content of the document entities in the document nodes, obtaining general descriptive text, and then converting the general descriptive text into vector values. When the document entity is relatively short, the entire document entity can also be directly converted into vector values.

[0024] In this embodiment, a document query request may involve one or more entities and the content of the desired document. When multiple entities are involved, the request may also include relationships between the entities.

[0025] For example, a document query request could be: "Please recommend some papers whose authors were students of the author of 'Research on XX Technology,' and whose research content is similar to the core ideas of this groundbreaking paper." This document query request includes paper entities, multiple author entities, and there is a writing relationship between the author entities and the paper entities, as well as a mentoring relationship between the author entities.

[0026] In this embodiment, when executing the aforementioned document query request, the document query request can first be analyzed to obtain the entities involved and the relationships between them. Then, graph retrieval is used to match multiple candidate document entities in the graph database. These candidate document entities can form a node candidate set, wherein the supervisor of the author of each candidate document entity is the author of "Research on XX Technology". Graph retrieval can use graph search algorithms, path discovery algorithms (such as Dijkstra's algorithm, A* algorithm, minimum spanning tree algorithm), etc. Graph retrieval excels at relation retrieval and can quickly, accurately, and efficiently match a node candidate set in the graph database.

[0027] The candidate node set may include many document nodes, some of which are highly relevant to "Research on XX Technology," while others are less relevant. Therefore, after obtaining the candidate node set, a semantic similarity search is needed to determine the few nodes with the highest similarity.

[0028] Specifically, the graph database can obtain the document query vector of the document query request, and the document query vector represents the content of "Research on XX Technology". In this embodiment, since the document query request contains the title of the desired document, the graph database can directly obtain the content attributes of "Research on XX Technology" based on its stored document nodes. In some embodiments, the document query request may not contain the title of the desired document, but only a summary of the desired document or individual paragraphs of text. In such cases, the graph database can convert the summary of the document query request or individual paragraphs of text into a vector form as the document query vector.

[0029] Next, the graph database retrieves the content attributes of each candidate document entity. For example, if the document node corresponding to a candidate document entity already stores content attributes in vector form, its vector values ​​can be read directly. If the document node corresponding to a candidate document entity does not store content attributes in vector form, they can be converted into vector-form content attributes based on its summary attributes.

[0030] Next, vector retrieval is used to match one or more of the content attributes of each candidate document entity that have the highest similarity to the document query vector, and these are selected as the target document entities. Vector retrieval can be performed using algorithms such as Inverted File with Flat (IVFFlat) and Hierarchical Navigable Small World (HNSW).

[0031] Vector retrieval excels at semantic analysis and semantic similarity search. By additionally setting the content attributes of each document node in the graph database in vector form and obtaining the document query vector of the document query request, the graph database can perform semantic search, thereby improving the accuracy of document query.

[0032] It's important to understand that vector retrieval requires significant memory and computational resources, and the larger the scope of the vector retrieval, the greater the overhead. In this embodiment, by first using graph retrieval to match a candidate set of nodes in the graph database, and then performing vector retrieval within that candidate set, the data range requiring vector retrieval is significantly reduced. The system does not need to perform vector similarity scanning on the entire dataset; instead, it performs calculations on a small, highly relevant subset, thereby greatly saving resource overhead.

[0033] The graph database data processing method in this embodiment, by using graph retrieval, quickly, accurately, and efficiently matches multiple candidate document entities in the graph database, significantly narrowing the data range that requires vector retrieval. By additionally setting vector-form content attributes for each document node in the graph database, and using vector retrieval to perform semantic similarity searches on the content attributes of multiple candidate document entities, semantics are accurately captured with minimal resource overhead, resulting in one or more target document entities with the highest similarity. Graph retrieval and vector retrieval support and cooperate with each other, achieving the goal of improving the overall accuracy and efficiency of document queries in the graph database.

[0034] In some embodiments of the graph database data processing method of the present invention, such as Figure 2 As shown, based on the document query request, graph retrieval is used to match multiple candidate document entities in the graph database, including: S211, retrieve the graph index of the entity nodes in the graph database, including document nodes and author nodes; S213, Obtain the structured relationships between multiple entities in the document query request, including document entities and author entities; S215, based on structured relationships and graph indexes, performs graph traversal operations to match multiple candidate document entities in the graph database.

[0035] Graph databases can create graph indexes during the node construction phase. Graph indexes can include attributes such as the title of document nodes and the name of author nodes, which can be set as needed. Using graph indexes for graph retrieval can improve retrieval efficiency.

[0036] In this embodiment, the system can perform multi-step reasoning based on the structured relationships between multiple entities in a document query request to obtain the corresponding graph path. Taking the document query request: "Please recommend some papers whose authors were students of the author of 'Research on XX Technology,' and whose research content is similar to the core ideas of this groundbreaking paper," as an example, the following graph path can be obtained: (:Paper)<-[:AUTHORED]-(author)-[:ADVISED]->(student)-[:AUTHORED]->(student_paper:Paper); Here, <- and -> indicate the direction of the relation edge.

[0037] Filtering based on graph paths can fully leverage the relational retrieval advantages of graph search, quickly, accurately, and efficiently matching the candidate set of nodes in the graph database, thereby narrowing the data range for subsequent vector retrieval.

[0038] In some embodiments of the graph database data processing method of the present invention, such as Figure 3 As shown, before the step of obtaining the document query vector of the document query request, the following steps are also included: S611, Construct the summary attributes of multiple document nodes in the graph database. The summary attributes are descriptive text used to describe the summary of the document entity of the corresponding document node. S613, Use the embedding model to obtain the embedding vectors of each summary attribute; S615: Construct the content attributes of each document node based on each embedding vector.

[0039] In this embodiment, the graph database reserves an attribute when constructing document nodes to store content attributes. When a document has a summary, the summary can be directly used as the summary attribute. When a document does not have a summary, manual or AI tools can be used to summarize the content of the document entities in the document node, obtain a summary descriptive text, and store the summary descriptive text as the summary attribute.

[0040] After obtaining the summary attributes, you can call the embedding model (such as the Embedding model) to convert the text of the summary attributes into vector form.

[0041] In this embodiment, document nodes store both summary attributes and content attributes. When semantic retrieval is not required, graph retrieval can be used to access the summary attributes; when semantic retrieval is needed, vector retrieval is used to access the content attributes. This solution innovatively combines graph model and vector storage to obtain a general document query process, providing a new solution for multi-model databases supporting both graph and vector types.

[0042] In some embodiments of the graph database data processing method of the present invention, such as Figure 3 As shown, after the steps of constructing the content attributes of each document node, the following steps are also included: S617, Construct vector indexes of the content attributes of document nodes in a graph database.

[0043] In this embodiment, the graph database creates a vector index of the content attributes of the document nodes while constructing them, so as to improve the retrieval efficiency when performing vector retrieval later.

[0044] In some embodiments of the graph database data processing method of the present invention, such as Figure 4 As shown, the document query vector for the document query request is obtained, including: S311, Obtain the title of the expected document in the document query request; S313, Match the document node corresponding to the title of the desired document in the graph database; S315, obtain the content attributes of the document node corresponding to the desired document, and use them as the document query vector.

[0045] When a document query request contains the title of the desired document, the corresponding document node can be matched from the title attribute of the graph database based on the title.

[0046] When the document nodes of each candidate document entity in the graph database have pre-stored content attributes, the document nodes corresponding to each candidate document entity can be directly read, and their content attributes can be read as document query vectors. When the content attributes are not pre-stored, an embedding model can be used to convert the summary attributes of the document nodes corresponding to each candidate document entity into vector form, which is then used as document query vectors and stored as the corresponding document node content attributes.

[0047] When a document query request does not contain the title of the desired document, but only a summary or individual paragraphs of the desired document, an embedding model can be used to convert the summary or individual paragraphs of the document query request into a vector form, which serves as the document query vector.

[0048] In some embodiments of the graph database data processing method of the present invention, such as Figure 5 As shown, using vector retrieval, the multiple candidate document entities with the highest similarity to the document query vector are matched from the content attributes of each candidate document entity and selected as the target document entities, including: S511, using vector indexing, calculate the similarity score between the content attributes of each candidate document entity and the document query vector; S513, sort the calculated similarity scores to obtain the N highest similarity scores, where N is an integer greater than 1; S515: Select the candidate document entities corresponding to the N highest similarity scores as the target document entities.

[0049] Using vector indexes can improve the efficiency of vector retrieval. In this embodiment, algorithms such as inverted file flattening and hierarchical navigable small-world graphs can be used to calculate the vector similarity score between each candidate document entity (e.g., a student paper) and the desired document (e.g., "Research on XX Technology"). The similarity score can be calculated using cosine similarity, Jaccard similarity, Euclidean distance, Manhattan distance, etc., but this is not a requirement. Because the use of graph retrieval in the early stages greatly narrows the data range required for vector retrieval, vector retrieval does not need to scan the entire dataset for vector similarity; instead, it is calculated on a small, highly relevant subset, significantly saving resource overhead.

[0050] After obtaining the similarity scores of each vector, the scores can be sorted in descending order, and the top N entities with the highest rankings can be returned as the final result.

[0051] In some embodiments of the graph database data processing method of the present invention, such as Figure 6 As shown, the methods for obtaining content attributes in vector form include: S711, when constructing each document node, an additional content attribute is reserved; S713, traverse all document nodes; S715, extract the summary attribute content of each document node; S717 calls the Embedding model to convert the summary text into vectors; S719, store the generated vector in the content attribute of the corresponding document node.

[0052] The flowchart provided in this embodiment is not intended to indicate that the operations of the method will be performed in any particular order, or that all operations of the method are included in every case. Furthermore, the method may include additional operations. Within the scope of the technical concept provided by the method in this embodiment, additional variations can be made to the above method.

[0053] It should be understood that in some embodiments, the components may be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods may be implemented using software or firmware stored in memory and executed by a suitable instruction execution system.

[0054] This invention also provides a computer program product 10, a computer-readable storage medium 20, and a computer device 30. Figure 7 This is a schematic diagram of a computer program product 10 according to an embodiment of the present invention. Figure 8 This is a schematic diagram of a computer-readable storage medium 20 according to an embodiment of the present invention. Figure 9 This is a schematic diagram of a computer device 30 according to an embodiment of the present invention. The computer program product 10 includes a computer program 11, which, when executed by the processor 32, implements the steps of any of the data processing methods described above. A computer-readable storage medium 20 stores the computer program 11 thereon, which, when executed by the processor 32, implements the steps of the data processing method of any of the embodiments described above. The computer device 30 may include a memory 31, a processor 32, and the computer program 11 stored in the memory 31 and running on the processor 32.

[0055] The computer program 11 used to perform the operations of this invention may be assembly instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, integrated circuit configuration data, or source code or object code written in any combination of one or more programming languages ​​and procedural programming languages. The computer program 11 may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer via any type of network, including a Local Area Network (LAN) or Wide Area Network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, to perform aspects of this invention, electronic circuits, including, for example, programmable logic circuits, Field-Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), may execute computer-readable program instructions to personalize the electronic circuits by utilizing state information from computer-readable program instructions.

[0056] For the purposes of this embodiment, computer program product 10 is a related product that includes computer program 11.

[0057] For the purposes of this embodiment, the computer-readable storage medium 20 is a tangible device capable of holding and storing a computer program 11. It can be any device capable of containing, storing, communicating, propagating, or transmitting the computer program 11 for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable storage medium 20 include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable optical disc read-only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanical encoding device, and any suitable combination thereof.

[0058] Computer device 30 can be, for example, a server, desktop computer, laptop computer, tablet computer, or smartphone. In some examples, computer device 30 can be a cloud computing node. Computer device 30 can be described in the general context of computer system executable instructions (such as program modules) executed by a computer system. Typically, program modules can include routines, programs, object programs, components, logic, data structures, etc., that perform specific tasks or implement specific abstract data types. Computer device 30 can be implemented in a distributed cloud computing environment where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules can reside on local or remote computing system storage media, including storage devices.

[0059] Computer device 30 may include a processor 32 adapted to execute stored instructions and a memory 31 that provides temporary storage space for the operation of said instructions during operation. The processor 32 may be a single-core processor, a multi-core processor, a computing cluster, or any other configuration. The memory 31 may include random access memory (RAM), read-only memory, flash memory, or any other suitable storage system.

[0060] Computer device 30 may also include a network adapter / interface and an input / output (I / O) interface. The I / O interface allows external devices that can be connected to the computer device to input and output data. The network adapter / interface provides communication between the computer device and a network, typically represented as a communication network.

[0061] Therefore, those skilled in the art should recognize that although numerous exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications conforming to the principles of the present invention can be directly determined or derived from the disclosure of the present invention without departing from the spirit and scope of the invention. Thus, the scope of the present invention should be understood and construed as covering all such other variations or modifications.

Claims

1. A data processing method for a graph database, characterized in that, include: Get the document query request; Based on the document query request, graph retrieval is used to match multiple candidate document entities in the graph database; Obtain the document query vector of the document query request, wherein the document query vector is used to characterize the content of the expected document of the document query request; Obtain the content attributes of each candidate document entity, wherein the content attributes are in vector form and are used to represent the content of the corresponding candidate document entity; Using vector retrieval, one or more of the content attributes of each candidate document entity are matched to the document query vector with the highest similarity, and these are selected as the target document entities.

2. The data processing method according to claim 1, characterized in that, The step of matching multiple candidate document entities in the graph database using graph retrieval based on the document query request includes: Obtain the graph index of the entity nodes in the graph database, where the entity nodes include document nodes and author nodes; Obtain the structured relationships between multiple entities in the document query request, where the entities include document entities and author entities; Based on the structured relationships and the graph index, a graph traversal operation is performed to match multiple candidate document entities in the graph database.

3. The data processing method according to claim 1, characterized in that, Before the step of obtaining the document query vector of the document query request, the method further includes: Construct summary attributes for multiple document nodes of the graph database. The summary attributes are descriptive text used to describe the summary of the document entity corresponding to the document node. Using an embedding model, the embedding vectors of each of the aforementioned summary attributes are obtained; Based on the embedding vectors, the content attributes of each document node are constructed.

4. The data processing method according to claim 3, characterized in that, The step of obtaining the document query vector of the document query request includes: Obtain the title of the desired document in the document query request; Match the document node in the graph database that corresponds to the title of the desired document; Obtain the content attributes of the document node corresponding to the desired document, and use them as the document query vector.

5. The data processing method according to claim 3, characterized in that, The step of obtaining the content attributes of each of the candidate document entities includes: The document nodes corresponding to each candidate document entity are obtained from the graph database, and their content attributes are read.

6. The data processing method according to claim 3, characterized in that, After the step of constructing the content attributes of each document node, the method further includes: Construct a vector index of the content attributes of the document nodes in the graph database.

7. The data processing method according to claim 6, characterized in that, The method of using vector retrieval, which matches the multiple candidate document entities with the highest similarity to the document query vector among their content attributes, and uses them as target document entities, includes: Using the vector index, calculate the similarity score between the content attribute of each candidate document entity and the document query vector; The calculated similarity scores are sorted to obtain the N highest similarity scores, where N is an integer greater than 1; The candidate document entities corresponding to the N highest similarity scores are taken as the target document entities.

8. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the steps of the data processing method for the graph database as described in any one of claims 1 to 7.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the data processing method for the graph database as described in any one of claims 1 to 7.

10. A computer device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the data processing method for the graph database according to any one of claims 1 to 7.

Images