Data model management system and method supporting multi-modal storage

Abstract

The invention discloses a data model management system and method supporting multi-modal storage. The method comprises the following steps: constructing a plurality of unified data models corresponding to different levels of an industry on a logic level according to business requirements and data analysis results; performing development processing on the unified data model of each level to establish a mutual relationship among the unified data models, and issuing the mutual relationship to take effect; a certain unified data model is selected to be applied to a specific data source type, and a corresponding relation with a physical model is established; converting the currently selected unified data model according to the specific data source type so as to enable the selected unified data model of the corresponding hierarchy to be matched with the specific data source type; a mapping relation between the selected unified data model and actual data storage is established, and a corresponding data source connector is provided during data access; and calling a corresponding data source connector according to the data source type to obtain physical object structure information, and executing a physical operation script to perform control operation on actual data. According to the method, association query across multiple data source types is realized in data model development.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a data model management system and method supporting multimodal storage. Background Art

[0002] In the process of data collection, processing, application and management, there is a need to process multiple data types. The data to be processed or processed exists in different data sources, such as relational databases, non-relational databases, big data components, etc. The same data type has unique storage characteristics in various data sources, such as data definition, data structure, data attributes, data operation syntax, etc., presenting a multimodal form; in traditional data operation processing, it is necessary to build a separate data model, develop a model, and implement a model based on each type of data source.

[0003] It is necessary to define, develop and build data models on different data sources for the same type of data, which results in high duplication of work, heavy workload and easy errors. At the same time, model development cannot support cross-data source type related queries, applications cannot be quickly migrated or switched or adapted to the cloud, and there are also many limitations in terms of support for information innovation transformation and model application. Summary of the invention

[0004] The purpose of the present invention is to provide a data model management system and method that supports multimodal storage. The data model processing semantics is separated from the specific data physical model. Through a unified data access engine, it can adapt to multiple types of databases and realize associative queries across multiple data source types in data model development.

[0005] The technical solution adopted by the present invention is: A data model management method supporting multimodal storage includes the following steps: Step 1: Build a number of unified data models corresponding to different levels of the industry at the logical level based on business needs and data analysis results, and define the data structure, data information, data features, data relationships, and data standards of the unified data models at different levels; Step 2: Develop and process unified data models at each level according to business requirements to establish relationships between unified data models, and publish them to make the corresponding unified data models formally effective; Step 3: Select a unified data model according to business needs and apply it to a specific data source type, and establish a corresponding relationship between the unified data model and the physical model; Step 4: convert the currently selected unified data model of the corresponding level according to the specific data source type, so that the selected unified data model of the corresponding level is adapted to the specific data source type; Step 5: According to the data information defined by the unified data model at the logical level and the specific data source type used, a mapping relationship between the selected unified data model and the actual data storage is established, and a corresponding data source connector is provided when accessing data; Step 6: Call the corresponding data source connector according to the data source type to obtain physical object structure information, and execute the physical operation script to implement specific control operations on the actual data.

[0006] Furthermore, when constructing a unified data model at the logical level in step 1, data structure, data information, data characteristics, data relations, and data standards are defined based on the logical level.

[0007] Furthermore, the specific steps of constructing a unified data model at the logical level in step 1 are as follows: Step 1-1, define the model hierarchy, classify the unified data model into several levels based on business needs and refer to the industry stratification method; and further divide the subject domain under each level as needed.

[0008] Step 1-2, model definition, defines the specific structure of the unified data model at different levels; Furthermore, set the model name, Chinese name, model level, data category, storage period, model applicability, and storage object naming constraints; set the business and technical calibers to describe the business rules, scenarios, and indicators of the data, and follow the principles of consistency, completeness, accuracy, and maintainability to provide guiding principles in data development.

[0009] Steps 1-3, logical field pool management, perform standard logical field management for unified data models at all levels according to industry standards, implement standardized naming and content settings of field information in the data model, preserve consistent semantics, and generate a public logical field pool; Steps 1-4 generate corresponding import rules for the unified data model at each level, so as to connect to the specific physical data source through the import rules to obtain the physical structure information of the specified table, realize the conversion from the physical model to the logical model, and generate the corresponding data model; Furthermore, the storage period of step 1-2 includes: permanent, day, week, month, year. When the value is set to "permanent", the data storage objects involved in the model save all data; when setting other options, you need to configure data deletion condition information. When the storage object data storage time exceeds the range, the storage object data is automatically cleared according to the configured condition information.

[0010] Furthermore, in steps 1-3, for the field information involved in the storage objects in the data model, field selection, loading, and application are provided through a public logical field pool so that it can be selected as a standard field by the data model; field attribute definition is provided, and field encoding, field name, field type, field length, field precision, primary key field, field attribute, data range, and data security settings are set; the data type is a logical definition including: integer, floating point, text, date and time, binary, etc.; data security provides data encryption and decryption, and data desensitization methods.

[0011] In addition to providing single field-level definition from the page, the logical field pool also provides the function of batch importing fields to the field pool through the physical model. Connect to the specified data source to extract the field information in the physical storage object structure, perform logical information conversion and load it into the field pool.

[0012] Furthermore, three development processing modes are provided in step 2: script development, model fusion, and model migration. The unified data model at the logical level is developed and processed by writing logic scripts, and the logic scripts do not involve specific physical syntax.

[0013] Specifically, in the script development mode, the logical semantic definition of data generation is implemented by writing data model scripts. This involves references to other data models, including model reading, association, and field selection. The data model script is a predefined method that is loaded and converted into a physical script for execution in the subsequent unified data access engine.

[0014] In the model fusion mode, through visual configuration, data models are referenced to implement multiple data model association queries. This mode provides model selection, field selection, data time setting, field calculation method, model association method, association primary key, association condition, and target field setting. The data time can be set as pseudo code, and the real value is replaced during the storage object data generation stage.

[0015] In the model migration mode, a designed data model is applied to a new field, and then the relevant information is adjusted according to the requirements, and finally a new data model is obtained.

[0016] After the data model is developed, it is published to make the data model officially effective.

[0017] Furthermore, in step 4, multiple types of parsers are provided to achieve automatic adaptation and conversion. The parsers include mysql parser, oracle parser, clickHouse parser, hive parser, doris parser, sqlServer parser, postreSql parser, etc.

[0018] Furthermore, the specific steps of step 4 are as follows: Step 4-1, data definition semantic conversion, applies the data structure, data attributes and other information defined in the data model to a specific data source type to generate an executable script, such as generating table creation statements, table structure modification statements, table deletion statements, etc.

[0019] Step 4-2, data operation language and logic development script conversion, according to the unified data model logic script syntax, adapt to the selected data source type to automatically convert and generate executable scripts. For cases involving multiple data source related queries, integration can be based on open source components such as trino.

[0020] Step 4-3, data model import operation, connect to the specified data source to obtain the specified table structure information, parse and generate data model related structure information, and realize the reverse conversion from physical model to logical model.

[0021] Furthermore, step 4-1 obtains the connection information of the data storage and performs a connectivity test. Instantiates the logical structure of the data model and builds the structural information of the data storage object. Obtain the field name, field type, field length, primary key, field description, value range, and nullability, and selects the available information to build the physical structure of the object.

[0022] In the unified data model logical field conversion phase, the logical field type needs to be converted to the field type of the actual physical table. When a logical field type corresponds to multiple field types of the physical table, a manual selection operation is provided.

[0023] Furthermore, the physical semantics of data generation processing in step 4-2 data model development are handled in two ways: using script development to obtain the data storage connection information corresponding to the model involved in the script, constructing table creation statements, and combining them with the data generation script in the data model to generate a complete executable script; and using page visualization selection to construct the relationship between the data models, and converting them into script mode at the bottom layer for processing.

[0024] Furthermore, in step 5, different connectors are provided for different data storage types to execute physical scripts. According to the structural information of the model, the data definition processing module converts it into table building operation semantics recognizable by the data storage for execution. According to the development script content in the data model, the data operation processing module is called to initiate a physical query. The development script in the data model to which the data storage object belongs can contain multiple source data models, and the source data model can be applied to different data storage types to realize cross-data source queries.

[0025] The present invention discloses a data model management system supporting multimodal storage, which includes a unified data model construction unit, a unified data model development unit, a unified data model application unit, a unified data model conversion unit, a unified data model bridging unit, a unified data access engine, and a data storage unit.

[0026] Unified data model building unit: abstractly defines the data structure, data information, data characteristics, data relationships, etc. carried by the business, and is created based on a logical method without involving specific data storage physical level content.

[0027] Unified data model development unit: implement data stratification and horizontal decoupling according to business needs and industry rules, define data processing and conversion processes at the logical level, provide data model hierarchical conversion, build new data models, and provide operational basis and processing constraints for subsequent applications.

[0028] Unified data model application unit: The data model is applied to the actual data source, providing automatic adaptation of data storage capabilities, realizing the conversion from logical semantics to physical semantics, and providing conditions for data model instantiation and data generation.

[0029] Unified data model conversion unit: Based on the structural definition and model development script of the unified data model, the logical script is converted according to the selected data source type to generate a recognizable and executable script for the corresponding data source type.

[0030] Unified data model bridging unit: used to establish associations between different data models, realize the intercommunication and conversion of data in the models. And based on the data model as an intermediary, it is applied to different types of databases to realize the rapid migration of data between databases.

[0031] Unified data access engine: Based on the generated execution script content, call the executor of the corresponding data source type for physical execution. Encapsulate the returned result data and generate data format and content that can be recognized by the unified data model.

[0032] Data storage unit: supports multiple data storage types for actual data storage. When the data model has a configured retention period, the data is automatically cleared.

[0033] The present invention adopts the above technical scheme, provides a unified data model management method to realize abstract data structure definition and data processing method, adapts to different types of data storage objects when the data model is applied, and constructs physical operation methods for corresponding storage objects; provides a unified data access engine based on the data processing method in the data model, supports multiple data source types in the model application, and provides query functions across data source types; realizes rapid migration of data storage objects based on the unified data model.

[0034] The structural definition and data processing semantics of the unified data model of the present invention can quickly adapt to objects of the same data type stored in different data. Business personnel, data managers, and data developers focus on the business-related content of the data model, can maintain consistent understanding of the meaning of the data, and can reduce repetitive operations. The data model processing semantics is separated from the specific data physical model. Through the unified data access engine, it can adapt to multiple types of databases, and realize cross-type data source related queries in data model development. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments; Figure 1 A schematic diagram of the principle architecture of a data model management method supporting multimodal storage according to the present invention. Implementation

[0036] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0037] like Figure 1 As shown, the present invention discloses a data model management method supporting multimodal storage, comprising the following steps: Step 1: Build a number of unified data models corresponding to different levels of the industry at the logical level based on business needs and data analysis results, and define the data structure, data information, data features, data relationships, and data standards of the unified data models at different levels; Step 2: Develop and process unified data models at each level according to business requirements to establish relationships between unified data models, and publish them to make the corresponding unified data models formally effective; Step 3: Select a unified data model according to business needs and apply it to a specific data source type, and establish a corresponding relationship between the unified data model and the physical model; Step 4: convert the currently selected unified data model of the corresponding level according to the specific data source type, so that the selected unified data model of the corresponding level is adapted to the specific data source type; Step 5: According to the data information defined by the unified data model at the logical level and the specific data source type used, a mapping relationship between the selected unified data model and the actual data storage is established, and a corresponding data source connector is provided when accessing data; Step 6: Call the corresponding data source connector according to the data source type to obtain physical object structure information, and execute the physical operation script to implement specific control operations on the actual data.

[0038] Furthermore, when constructing a unified data model at the logical level in step 1, data structure, data information, data characteristics, data relations, and data standards are defined based on the logical level.

[0039] Furthermore, the specific steps of constructing a unified data model at the logical level in step 1 are as follows: Step 1-1, define the model hierarchy, classify the unified data model into several levels based on business needs and refer to the industry stratification method; and further divide the subject domain under each level as needed.

[0040] Step 1-2, model definition, defines the specific structure of the unified data model at different levels; Furthermore, set the model name, Chinese name, model level, data category, storage period, model applicability, and storage object naming constraints; set the business and technical calibers to describe the business rules, scenarios, and indicators of the data, and follow the principles of consistency, completeness, accuracy, and maintainability to provide guiding principles in data development.

[0041] Steps 1-3, logical field pool management, perform standard logical field management for unified data models at all levels according to industry standards, implement standardized naming and content settings of field information in the data model, preserve consistent semantics, and generate a public logical field pool; Steps 1-4 generate corresponding import rules for the unified data model at each level, so as to connect to the specific physical data source through the import rules to obtain the physical structure information of the specified table, realize the conversion from the physical model to the logical model, and generate the corresponding data model; Furthermore, the storage period of step 1-2 includes: permanent, day, week, month, year. When the value is set to "permanent", the data storage objects involved in the model save all data; when setting other options, you need to configure data deletion condition information. When the storage object data storage time exceeds the range, the storage object data is automatically cleared according to the configured condition information.

[0042] Furthermore, in steps 1-3, for the field information involved in the storage objects in the data model, field selection, loading, and application are provided through a public logical field pool so that it can be selected as a standard field by the data model; field attribute definition is provided, and field encoding, field name, field type, field length, field precision, primary key field, field attribute, data range, and data security settings are set; the data type is a logical definition including: integer, floating point, text, date and time, binary, etc.; data security provides data encryption and decryption, data desensitization methods, etc.

[0043] In addition to providing single field-level definition from the page, the logical field pool also provides the function of batch importing fields to the field pool through the physical model. Connect to the specified data source to extract the field information in the physical storage object structure, perform logical information conversion and load it into the field pool.

[0044] Furthermore, three development processing modes are provided in step 2: script development, model fusion, and model migration. The unified data model at the logical level is developed and processed by writing logic scripts, and the logic scripts do not involve specific physical syntax.

[0045] Specifically, in the script development mode, the logical semantic definition of data generation is implemented by writing data model scripts. This involves references to other data models, including model reading, association, and field selection. The data model script is a predefined method that is loaded and converted into a physical script for execution in the subsequent unified data access engine.

[0046] In the model fusion mode, through visual configuration, data models are referenced to implement multiple data model association queries. This mode provides model selection, field selection, data time setting, field calculation method, model association method, association primary key, association condition, and target field setting. The data time can be set as pseudo code, and the real value is replaced during the storage object data generation stage.

[0047] In the model migration mode, a designed data model is applied to a new field, and then the relevant information is adjusted according to the requirements, and finally a new data model is obtained.

[0048] After the data model is developed, it is published to make the data model officially effective.

[0049] Furthermore, in step 4, multiple types of parsers are provided to achieve automatic adaptation and conversion. The parsers include mysql parser, oracle parser, clickHouse parser, hive parser, doris parser, sqlServer parser, postreSql parser, etc.

[0050] Furthermore, the specific steps of step 4 are as follows: Step 4-1, data definition semantic conversion, applies the data structure, data attributes and other information defined in the data model to a specific data source type to generate an executable script, such as generating table creation statements, table structure modification statements, table deletion statements, etc.

[0051] Step 4-2, data operation language and logic development script conversion, according to the unified data model logic script syntax, adapt to the selected data source type to automatically convert and generate executable scripts. For cases involving multiple data source related queries, integration can be based on open source components such as trino.

[0052] Step 4-3, data model import operation, connect to the specified data source to obtain the specified table structure information, parse and generate data model related structure information, and realize the reverse conversion from physical model to logical model.

[0053] Furthermore, step 4-1 obtains the connection information of the data storage and performs a connectivity test. Instantiates the logical structure of the data model and builds the structural information of the data storage object. Obtain the field name, field type, field length, primary key, field description, value range, and nullability, and selects the available information to build the physical structure of the object.

[0054] In the unified data model logical field conversion phase, the logical field type needs to be converted to the field type of the actual physical table. When a logical field type corresponds to multiple field types of the physical table, a manual selection operation is provided.

[0055] Furthermore, the physical semantics of data generation processing in step 4-2 data model development are handled in two ways: using script development to obtain the data storage connection information corresponding to the model involved in the script, constructing table creation statements, and combining them with the data generation script in the data model to generate a complete executable script; and using page visualization selection to construct the relationship between the data models, and converting them into script mode at the bottom layer for processing.

[0056] Furthermore, in step 5, different connectors are provided for different data storage types to execute physical scripts. According to the structural information of the model, the data definition processing module converts it into table building operation semantics recognizable by the data storage for execution. According to the development script content in the data model, the data operation processing module is called to initiate a physical query. The development script in the data model to which the data storage object belongs can contain multiple source data models, and the source data model can be applied to different data storage types to realize cross-data source queries.

[0057] Figure 1 As shown, the present invention discloses a data model management system supporting multimodal storage, which includes unified data model construction, unified data model development, unified data model application, unified data model conversion, unified data model bridging, unified data access engine, and data storage module.

[0058] Unified data model construction: abstractly define the data structure, data information, data characteristics, data relationships, etc. carried by the business, create it based on logic, and do not involve the content of the specific physical level of data storage.

[0059] Unified data model development: Implement data stratification and horizontal decoupling based on business needs and industry rules, define data processing and conversion processes at the logical level, provide data model hierarchical conversion, build new data models, and provide operational basis and processing constraints for subsequent applications.

[0060] Unified data model application: The data model is applied to the actual data source, providing automatic adaptation of data storage capabilities, realizing the conversion from logical semantics to physical semantics, and providing conditions for data model instantiation and data generation.

[0061] Unified data model conversion: Based on the structural definition and model development script of the unified data model, perform logical script conversion according to the selected data source type to generate identifiable and executable scripts for the corresponding data source type.

[0062] Unified data model bridging: used to establish associations between different data models, realize the intercommunication and conversion of data in the models. And based on the data model as an intermediary, it is applied to different types of databases to realize the rapid migration of data between databases.

[0063] Unified data access engine: Based on the generated execution script content, call the executor of the corresponding data source type for physical execution. Encapsulate the returned result data and generate data format and content that can be recognized by the unified data model.

[0064] Data storage: Supports multiple data storage types for actual data storage. When the data model has a configured retention period, data clearing is automatically performed.

[0065] The present invention adopts the above technical scheme, provides a unified data model management method to realize abstract data structure definition and data processing method, adapts to different types of data storage objects when the data model is applied, and constructs physical operation methods for corresponding storage objects; provides a unified data access engine based on the data processing method in the data model, supports multiple data source types in the model application, and provides query functions across data source types; realizes rapid migration of data storage objects based on the unified data model.

[0066] The structural definition and data processing semantics of the unified data model of the present invention can quickly adapt to objects of the same data type stored in different data. Business personnel, data managers, and data developers focus on the business-related content of the data model, can maintain consistent understanding of the meaning of the data, and can reduce repetitive operations. The data model processing semantics is separated from the specific data physical model. Through the unified data access engine, it can adapt to multiple types of databases, and realize cross-type data source related queries in data model development.

[0067] Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians of the art without making creative work are within the scope of protection of the present application.

Claims

1. A data model management method supporting multimodal storage, characterized in that: It includes the following steps: Step 1: Build a number of unified data models corresponding to different levels of the industry at the logical level based on business needs and data analysis results, and define the data structure, data information, data features, data relationships, and data standards of the unified data models at different levels; Step 2: Develop and process unified data models at each level according to business requirements to establish relationships between unified data models, and publish them to make the corresponding unified data models formally effective; Step 3: Select a unified data model according to business needs and apply it to a specific data source type, and establish a corresponding relationship between the unified data model and the physical model; Step 4: convert the currently selected unified data model of the corresponding level according to the specific data source type, so that the selected unified data model of the corresponding level is adapted to the specific data source type; Step 5: According to the data information defined by the unified data model at the logical level and the specific data source type used, a mapping relationship between the selected unified data model and the actual data storage is established, and a corresponding data source connector is provided when accessing data; Step 6: Call the corresponding data source connector according to the data source type to obtain physical object structure information, and execute the physical operation script to implement specific control operations on the actual data.

2. The data model management method supporting multimodal storage according to claim 1, characterized in that: Step 1: Build a unified data model at the logical level. The specific steps are as follows: Step 1-1, define model classification: classify the unified data model into several levels according to business needs and refer to the industry stratification method; and further divide the subject domain under each level as needed; Step 1-2, model definition: define the specific structure of the unified data model at different levels; Step 1-3, logical field pool management: perform standard logical field management for unified data models at all levels according to industry standards, implement standardized naming and content settings of field information in the data model, preserve consistent semantics, and generate a public logical field pool; Steps 1-4, generate import rules: Generate corresponding import rules for the unified data model at each level, so as to connect to the specific physical data source through the import rules to obtain the physical structure information of the specified table, realize the conversion from the physical model to the logical model, and generate the corresponding data model.

3. The data model management method supporting multimodal storage according to claim 2, characterized in that: Steps 1-2 specifically include setting the model name, Chinese name, model level, data category, storage period, model application scope, and storage object naming constraints; setting business and technical calibers to describe the business rules, scenarios, and indicators of the data; the storage period includes: permanent, day, week, month, and year; when the value is set to "permanent", the data storage objects involved in the model save all data; when setting other options, you need to configure data deletion condition information. When the storage object data storage time exceeds the range, the storage object data is automatically cleared according to the configured condition information; In steps 1-3, for the field information involved in the storage object in the data model, the public logical field pool provides field selection, loading, and application, so that it can be selected as a standard field by the data model; Provide field attribute definition, set field encoding, field name, field type, field length, field precision, primary key field, field attribute, data range, data security settings; data types are logically defined and include: integer, floating point, text, date and time, binary, etc.; data security provides data encryption and decryption, and data desensitization methods; In addition to providing single field-level definition from the page, the logical field pool also provides the function of batch importing fields into the field pool through the physical model; connecting to the specified data source to extract the field information in the physical storage object structure, performing logical information conversion and loading it into the field pool.

4. The data model management method supporting multimodal storage according to claim 1, characterized in that: Step 2 provides three development processing modes: script development, model fusion, and model migration; Among them, in the script development mode, the logical semantic definition of data generation is realized by writing data model scripts; it involves references to other data models, including model reading, association, and field selection; the data model script is a predefined method, which is loaded and converted into a physical script for execution in the subsequent unified data access engine; In the model fusion mode, data models are referenced through visual configuration to realize multiple data model association queries; the model fusion mode provides model selection, field selection, data time setting, field calculation method, model association method, association primary key, association condition, and target field setting; the data time setting pseudo code replaces the real value in the storage object data generation stage; In the model migration mode, a designed data model is applied to a new domain, and then the information is adjusted according to the requirements, and finally a new data model is obtained; After the data model is developed, it is published to make the data model officially effective.

5. The data model management method supporting multimodal storage according to claim 1, characterized in that: In step 4, multiple types of parsers are provided to achieve automatic adaptation and conversion; the parsers include mysql parser, oracle parser, clickHouse parser, hive parser, doris parser, sqlServer parser, postreSql parser, etc.

6. The data model management method supporting multimodal storage according to claim 1, characterized in that: The specific steps of step 4 are as follows: Step 4-1, data definition semantic conversion: Generate an executable script based on the data structure and data attribute information defined in the data model and applied to a specific data source type; Step 4-2, data operation language and logic development script conversion: automatically convert and generate executable scripts according to the logic script syntax in the unified data model and the selected data source type; Step 4-3, data model import operation: connect to the specified data source to obtain the specified table structure information, parse and generate the structure information of the data model, and realize the reverse conversion from the physical model to the logical model.

7. The data model management method supporting multimodal storage according to claim 6, characterized in that: Step 4-1 obtains the connection information of the data storage to perform connectivity testing; instantiates the logical structure of the data model to construct the structural information of the data storage object; Get the field name, field type, field length, primary key, field description, value range, nullable status, and select the available information to build the physical structure of the object; During the unified data model logical field conversion phase, the logical field type is converted to the field type of the actual physical table; when a logical field type corresponds to multiple field types of the physical table, a manual selection operation is provided.

8. The data model management method supporting multimodal storage according to claim 6, characterized in that: Step 4-2 In data model development, the physical semantics of data generation processing are processed in two ways: using script development to obtain the data storage connection information corresponding to the model involved in the script to construct a table creation statement, and then combine it with the data generation script in the data model to generate a complete executable script; The relationship of the data model is constructed by means of page visualization selection, and is uniformly converted into script mode for processing at the bottom layer.

9. The data model management method supporting multimodal storage according to claim 1, characterized in that: In step 5, different connectors are provided for different data storage types to execute physical scripts; the data definition processing module converts the structural information of the model into table building operation semantics that can be recognized by the data storage for execution; and the data operation processing module is called to initiate physical query according to the development script content in the data model; The development script in the data model to which the data storage object belongs contains multiple source data models, and the source data models are applied to different data storage types to implement cross-data source queries.

10. A data model management system supporting multimodal storage, applying a data model management method supporting multimodal storage as claimed in any one of claims 1 to 9, characterized in that: The system includes a unified data model construction unit, a unified data model development unit, a unified data model application unit, a unified data model conversion unit, a unified data model bridge unit, a unified data access engine, and a data storage unit; Unified data model building unit: used to abstractly define the data structure, data information, data characteristics, and data relationships carried by the business. It is created based on logic and does not involve the content of the physical level of specific data storage; Unified data model development unit: used to implement data stratification and horizontal decoupling according to business needs and industry rules, define data processing and conversion processes at the logical level, provide data model hierarchical conversion, build new data models, and provide operation basis and processing constraints for subsequent applications; Unified data model application unit: used to apply the data model to the actual data source, provide automatic adaptation data storage capabilities, realize the conversion from logical semantics to physical semantics, and provide conditions for data model instantiation and data generation; Unified data model conversion unit: used to develop scripts based on the structure definition and model of the unified data model, perform logic script conversion according to the selected data source type, and generate identifiable and executable scripts for the corresponding data source type; Unified data model bridging unit: used to establish associations between different data models, realize the intercommunication and conversion of data in the models, and use the data model as an intermediary to apply to different types of databases to achieve rapid data migration between databases; Unified data access engine: Based on the generated execution script content, call the executor of the corresponding data source type for physical execution; Encapsulate the returned result data to generate data format and content that can be recognized by the unified data model; Data storage unit: supports multiple data storage types for actual data storage; When the data model has a configured retention period, data clearing operations are automatically performed.

Images