A server-based data management system

Abstract

The invention discloses a data management system based on a server, relates to the technical field of data management, and is used to improve the problem of low storage efficiency caused by structural similarity in stored data. The invention comprises the following steps: collecting keyword prefixes and suffixes of file names of stored data and transmission time intervals, using a geometric mean method to preliminarily classify the stored data and preselect a storage method, detecting the standard processing algorithms of all data files in the stored data and the joint call frequency between data, using a polynomial regression method to judge whether the stored data is to be split and stored, selecting a structural storage method for storage or data splitting according to the judgment result, marking each data file in the stored data when performing data splitting, obtaining the file update frequency and storage space occupied by each marked data file, using a hierarchical analysis method to screen the marked data files, and selecting a non-structural storage method for storage of the screened marked data files.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and more particularly to a server-based data management system. Background Art

[0002] Data management technology is a method and tool used to collect, store, process and analyze data, aiming to ensure the validity and security of data. The application of data management technology in servers can make data interacting through the server more convenient and secure.

[0003] The prior art has the following deficiencies:

[0004] In the past, when storing data in a server, structured storage or unstructured storage was used according to the data content. However, most stored data contains a large number of data files with similar structures and a small number of other data files. Using structured storage or unstructured storage for such stored data will cause inconvenience in subsequent calls to the stored data, reducing storage efficiency and security. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a server-based data management system, which preliminarily selects a storage method by analyzing the structured differences of each data file in the stored data, further detects the stored data, and determines which data files in the stored data need to be separately stored in an unstructured manner to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A server-based data management system includes a data acquisition module, a structure classification module, a splitting judgment module and a storage switching module, and each module is signal-connected;

[0008] The data collection module is used to collect the keyword prefixes and suffixes of the file names and the transmission time interval of the data files in the storage data and pass them to the structure classification module. After receiving the classification results passed in by the structure classification module, the standard processing algorithm of all data files in the storage data and the joint call frequency are detected and passed to the split judgment module;

[0009] The structural classification module classifies the stored data according to the keyword prefix and suffix of the file name of the data file in the stored data and the transmission time interval, and transmits the classification result back to the data acquisition module;

[0010] The split judgment module analyzes the maximum processing ratio of the same standard processing algorithm according to the standard processing algorithm of all data files in the storage data, comprehensively considers the maximum processing ratio of the storage data and the joint call frequency to determine whether to pass the storage data into the split storage mechanism, and transmits the judgment result to the storage switching module;

[0011] The storage switching module selects the structured storage method for storage or data splitting based on the judgment result. Before data splitting, it marks each data file in the storage data, obtains the file update frequency and storage space occupied by each marked data file, filters the marked data files, and selects the non-structured storage method for storage of the filtered marked data files.

[0012] In a preferred embodiment, the data acquisition module obtains the keyword prefixes and suffixes of the file names of each data file in the stored data, counts the types of the prefixes and suffixes, obtains the number of keyword prefixes of various types of file names and the number of keyword suffixes of various types of file names, and merges them into a prefix and suffix data set, uses the maximum value selected from the prefix and suffix data set to calculate the ratio with the total number of all data files in the stored data, and uses the calculation result as the prefix and suffix duplication percentage;

[0013] The average of the transmission time intervals of different data files in the stored data is taken as the mean transmission time. The data acquisition module passes the prefix and suffix duplication percentage of the stored data and the mean transmission time to the structure classification module.

[0014] In a preferred embodiment, after receiving the prefix and suffix repetition percentage and the transmission time mean of the stored data, the structure classification module uses the geometric mean method to classify the stored data into high structure data or low structure data. The specific steps are as follows:

[0015] Data preprocessing: The mean transmission time of the stored data is normalized using logarithmic changes, using the logarithmic change formula: y=ln x, where x is the mean transmission time of the stored data, and y is the normalized result of the mean transmission time of the stored data;

[0016] Geometric mean analysis: The geometric mean formula is constructed by normalizing the percentage of prefix and suffix duplication of the stored data with the mean of the transmission time: , where y is the normalized result of the mean value of the transmission time of the stored data, z is the percentage of prefix and suffix repetition of the stored data, and L is the result of the geometric mean formula, which is used as the data classification coefficient;

[0017] Data classification: When the data classification coefficient exceeds the preset classification threshold, the stored data will be classified as high-structure data and the structural storage method will be selected for storage; when the data classification coefficient is lower than the classification threshold, the stored data will be classified as low-structure data and transferred to the data acquisition module.

[0018] In a preferred embodiment, after receiving the classification results transmitted by the structure classification module, the data acquisition module accesses the log records to detect the standard processing algorithms of all data files in the stored data and the joint call frequency between the data;

[0019] The data acquisition module selects a period of time as the analysis time, analyzes each data file in the storage data within the analysis time, counts the ratio of the number of joint calls of each data file in the storage data to the total number of calls as the call frequency of the corresponding data file, and selects the median from the call frequencies of all data files as the joint call frequency of the storage data;

[0020] The joint call is to call other data files on the basis of calling the current data file in the storage data within a preset time interval.

[0021] In a preferred embodiment, the split judgment module receives the standard processing algorithm of all data files in the stored data and then calculates and obtains the maximum processing ratio of the stored data as follows:

[0022] The split judgment module sorts all data files in the storage data, starting from the standard processing algorithm of the first data file in the storage data, records its category and assigns a value of 1, and after the assignment is completed, traverses the subsequent data files one by one according to the sorting order. When the standard processing algorithm of the subsequent data file is the same as the standard processing algorithm of the assigned data file, the assignment on the corresponding category is increased by 1. When the standard processing algorithm of the subsequent data file is different from the standard processing algorithm of the assigned data file, its category is recorded and assigned a value of 1 until all data files in the storage data are traversed;

[0023] The category with the largest assigned value is selected from the stored data and recorded as the largest category. The ratio of the assigned value of the largest category to the total assigned values ​​of the categories of all data files in the stored data is taken as the maximum processing ratio of the stored data.

[0024] In a preferred embodiment, the split judgment module comprehensively analyzes the stored data by using the polynomial regression method based on the maximum processing ratio of the stored data and the joint call frequency, and determines whether to import the stored data into the split storage mechanism. The specific process is as follows:

[0025] The maximum processing ratio of stored data and the joint call frequency are used to construct a polynomial regression formula to calculate the regression results: , where a is the maximum processing ratio of stored data, b is the frequency of joint calls of stored data, and c is the reconciliation parameter. and are the parameter weights of a and b respectively, and D is the regression result;

[0026] The regression result obtained by calculating the maximum processing ratio of the stored data and the joint call frequency is compared with the preset judgment threshold. When the regression result exceeds the judgment threshold, the judgment result is set to 0; otherwise, the data result is set to 1.

[0027] In a preferred embodiment, the storage switching module performs the following processing according to the judgment result transmitted by the split judgment module:

[0028] When the judgment result is 0, the data is not split and the stored data is stored using the structured storage method;

[0029] When the judgment structure is 1, the data is split and each data file in the stored data is marked differently;

[0030] The storage switching module selects a period of time as the update time, records the number of file updates of each marked data file within the update time, uses the number of file updates of the marked data file as the file update frequency of the corresponding marked data file, records the occupied space after each update of the marked data file within the update time and calculates the average value as the occupied space of the corresponding marked data file.

[0031] In a preferred embodiment, the storage switching module comprehensively stores the file update frequency and storage space of each marked data file in the data using the hierarchical analysis method to set the storage weight of each marked data file, and screens the marked data files according to the storage weight. The specific steps are as follows:

[0032] Preliminary data processing: standardize the space occupied by each labeled data file;

[0033] Calculate the weight reference: sum the file update frequency of each marked data file and the occupied space after normalization, and use the sum result as the weight reference;

[0034] Select a point scale: Select a point scale from 1 to 4, where 1 indicates an equal importance difference, 2 indicates a slight importance difference, 3 indicates a moderate importance difference, and 4 indicates a severe importance difference;

[0035] Construct feature matrix: construct feature matrix by setting point quantity according to weight reference quantity of each labeled data file;

[0036] Simplify the feature matrix: normalize each column of the feature matrix so that the sum of each column is 1;

[0037] Calculate feature vector: Calculate the feature vector of each labeled data file as the storage weight.

[0038] In a preferred embodiment, the storage switching module uses the average storage weight of all marked data files in the storage data as the split threshold; screens out the marked data files in the storage data whose storage weight exceeds the split threshold and selects the non-structural storage method for storage.

[0039] Technical effects and advantages of a server-based data management system of the present invention:

[0040] The present invention performs preliminary classification of stored data by collecting keyword prefixes and suffixes of file names and transmission time intervals of stored data, pre-selects storage methods according to classification results, narrows the data range, reduces splitting costs, detects standard processing algorithms of all data files in the stored data and the frequency of joint calls between data, determines whether the stored data is to be split and stored, improves the splitting direction and accuracy through one judgment, selects a structural storage method for storage or data splitting according to the judgment result, marks each data file in the stored data when performing data splitting, obtains the file update frequency and storage space occupied by each marked data file, screens the marked data files, selects a non-structural storage method for storage of the screened marked data files, and improves the storage efficiency and storage security of the stored data. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic diagram of a server-based data management system of the present invention.

[0042] Figure 2 The present invention is a flow chart of a server-based data management system. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] The present invention performs preliminary classification of stored data by collecting keyword prefixes and suffixes of file names and transmission time intervals of stored data, pre-selects storage methods according to classification results, detects standard processing algorithms of all data files in the stored data and the frequency of joint calls between data, determines whether the stored data is split for storage, selects a structural storage method for storage or data splitting according to the determination result, marks each data file in the stored data when performing data splitting, obtains the file update frequency and storage space occupied by each marked data file, screens the marked data files, selects a non-structural storage method for storage of the screened marked data files, and improves the storage efficiency and storage security of the stored data.

[0045] Embodiment, a server-based data management system, such as Figure 1 As shown, it includes a data acquisition module, a structure classification module, a splitting judgment module and a storage switching module, and the signals between the modules are connected;

[0046] The functions of each module are as follows:

[0047] The data collection module is used to collect the keyword prefixes and suffixes of the file names and the transmission time interval of the data files in the storage data and pass them to the structure classification module. After receiving the classification results passed in by the structure classification module, the standard processing algorithm of all data files in the storage data and the joint call frequency are detected and passed to the split judgment module;

[0048] The structural classification module classifies the stored data according to the keyword prefix and suffix of the file name of the data file in the stored data and the transmission time interval, and transmits the classification result back to the data acquisition module;

[0049] The split judgment module analyzes the maximum processing ratio of the same standard processing algorithm according to the standard processing algorithm of all data files in the storage data, comprehensively considers the maximum processing ratio of the storage data and the joint call frequency to determine whether to pass the storage data into the split storage mechanism, and transmits the judgment result to the storage switching module;

[0050] The storage switching module selects the structured storage method for storage or data splitting based on the judgment result. Before data splitting, it marks each data file in the storage data, obtains the file update frequency and storage space occupied by each marked data file, filters the marked data files, and selects the non-structured storage method for storage of the filtered marked data files.

[0051] It should be noted that the stored data is a data packet that needs to be stored. The stored data contains multiple data files. The standard processing method is a series of steps and procedures to ensure that the processing of data, information or processes complies with specific regulations. It can be obtained through access log records. The higher the correlation of data files with the same standard processing method in the stored data, the more likely it is that the use of structured storage methods will facilitate subsequent call analysis.

[0052] The data acquisition module performs SQL query on the data files in the storage data to obtain the prefix and suffix of the file name keyword of each data file, counts the types of prefix and suffix, obtains the number of various types of file name keyword prefixes and the number of various types of file name keyword suffixes, and merges them into a prefix and suffix data set, uses the maximum value selected from the prefix and suffix data set to calculate the ratio with the total number of all data files in the storage data, and uses the calculation result as the prefix and suffix duplication percentage;

[0053] The average of the transmission time intervals of different data files in the stored data is taken as the mean transmission time. The data acquisition module passes the prefix and suffix duplication percentage of the stored data and the mean transmission time to the structure classification module.

[0054] It should be noted that SQL query is a language that interacts with relational database management systems. In this example, it is used to check the keyword prefixes and suffixes of the file names of the data files in the stored data, perform quantity statistics, and merge the prefix and suffix data sets. The percentage of prefix and suffix duplication of the stored data is presented in decimal form, which is convenient for subsequent analysis and calculation when classifying the stored data.

[0055] After receiving the prefix and suffix duplication percentage and the mean transmission time of the stored data, the structure classification module uses the geometric mean method to classify the stored data into high structure data or low structure data. The specific steps are as follows:

[0056] Data preprocessing: The mean transmission time of the stored data is normalized using logarithmic changes, using the logarithmic change formula: y=ln x, where x is the mean transmission time of the stored data, and y is the normalized result of the mean transmission time of the stored data;

[0057] Geometric mean analysis: The geometric mean formula is constructed by normalizing the percentage of prefix and suffix duplication of the stored data with the mean of the transmission time: , where y is the normalized result of the mean value of the transmission time of the stored data, z is the percentage of prefix and suffix repetition of the stored data, and L is the result of the geometric mean formula, which is used as the data classification coefficient;

[0058] Data classification: When the data classification coefficient exceeds the preset classification threshold, the stored data will be classified as high-structure data and the structural storage method will be selected for storage; when the data classification coefficient is lower than the classification threshold, the stored data will be classified as low-structure data and transferred to the data acquisition module.

[0059] It should be noted that the classification threshold can be set according to actual conditions, for example, the classification threshold can be set to 0.6, etc., which will not be elaborated here.

[0060] After receiving the classification results transmitted by the structure classification module, the data acquisition module accesses the log records to detect the standard processing algorithms of all data files in the stored data and the joint call frequency between data;

[0061] The data acquisition module selects a period of time as the analysis time, analyzes each data file in the storage data within the analysis time, counts the ratio of the number of joint calls of each data file in the storage data to the total number of calls as the call frequency of the corresponding data file, and selects the median from the call frequencies of all data files as the joint call frequency of the storage data;

[0062] The joint call is to call other data files on the basis of calling the current data file in the storage data within a preset time interval.

[0063] The data acquisition module sends the maximum processing ratio of the stored data and the joint call frequency to the split judgment module for split judgment.

[0064] It should be noted that the preset time interval for the joint call analysis can be set by yourself, for example, the time interval can be set to 20s, etc., which will not be analyzed in detail here.

[0065] After receiving the standard processing algorithm of all data files in the storage data, the split judgment module calculates and obtains the maximum processing ratio of the storage data. The process is as follows:

[0066] The split judgment module sorts all data files in the stored data, starting from the standard processing algorithm of the first data file in the stored data, records its category and assigns a value of 1. After the assignment is completed, the subsequent data files are traversed one by one according to the sorting order. When the standard processing algorithm of the subsequent data file is the same as the standard processing algorithm of the assigned data file, the assignment on the corresponding category is increased by 1. When the standard processing algorithm of the subsequent data file is different from the standard processing algorithm of the assigned data file, its category is recorded and assigned a value of 1 until all data files in the stored data are traversed.

[0067] The category with the largest assigned value is selected from the stored data and recorded as the largest category. The ratio of the assigned value of the largest category to the total assigned values ​​of the categories of all data files in the stored data is taken as the maximum processing ratio of the stored data.

[0068] The split judgment module uses the polynomial regression method to analyze the stored data based on the maximum processing ratio of the stored data and the joint call frequency, and determines whether to import the stored data into the split storage mechanism. The specific process is as follows:

[0069] The maximum processing ratio of stored data and the joint call frequency are used to construct a polynomial regression formula to calculate the regression results: , where a is the maximum processing ratio of stored data, b is the frequency of joint calls of stored data, and c is the reconciliation parameter. and are the parameter weights of a and b respectively, and D is the regression result;

[0070] The regression result obtained by calculating the maximum processing ratio of the stored data and the joint call frequency is compared with the preset judgment threshold. When the regression result exceeds the judgment threshold, the judgment result is set to 0; otherwise, the data result is set to 1.

[0071] The split judgment module transmits the judgment result to the storage switching module, and the storage switching module performs selection processing according to the judgment result.

[0072] It should be noted that the parameter weights in the polynomial regression formula are set by professionals in this field, and the harmonization parameters are used to adjust the regression results to a specified judgment interval. The setting results are not unique, and the control parameters can be set to 0.1, etc.

[0073] The storage switching module processes the judgment result transmitted by the split judgment module as follows:

[0074] When the judgment result is 0, the data is not split and the stored data is stored using the structured storage method;

[0075] When the judgment structure is 1, data splitting is performed, and different marks are given to each data file in the stored data.

[0076] The storage switching module selects a period of time as the update period, records the number of file updates of each marked data file within the update period, uses the number of file updates of the marked data file as the file update frequency of the corresponding marked data file, and records the occupied space of the marked data file after each update within the update period and calculates the average value as the occupied space of the corresponding marked data file;

[0077] The file update frequency and storage space occupied by each marked data file in the comprehensive storage data are used to set the storage weight of each marked data file using the hierarchical analysis method, and the marked data files are screened according to the storage weight. The specific steps are as follows:

[0078] Preliminary data processing: The Max-Min standardization method is used to standardize the storage space occupied by each labeled data file;

[0079] Calculate the weight reference: sum the file update frequency of each marked data file and the occupied space after normalization, and use the sum result as the weight reference;

[0080] Select a point scale: Select a point scale from 1 to 4, where 1 indicates an equal importance difference, 2 indicates a slight importance difference, 3 indicates a moderate importance difference, and 4 indicates a severe importance difference;

[0081] Construct feature matrix: Construct feature matrix for all labeled data files in the storage data, as shown below:

[0082] Simplify the feature matrix: Normalize each column of the feature matrix so that the sum of each column is 1, as shown below:

[0083] Calculate the feature vector: Calculate the feature vector of each labeled data file. The feature vector is the weighted average of each column of the feature matrix. The feature vector in the feature matrix is ​​used as the storage weight of each labeled data file in the feature matrix.

[0084] A tag data file storage weight = (0.1000 + 0.0770 + 0.0870 + 0.1200) / 4 = 0.0960;

[0085] B tag data file storage weight = (0.2000 + 0.1538 + 0.1305 + 0.1600) / 4 = 0.1611;

[0086] C-marked data file storage weight = (0.3000 + 0.3077 + 0.2608 + 0.2400) / 4 = 0.2771;

[0087] D marked data file storage weight = (0.4000 + 0.4615 + 0.5217 + 0.4800) / 4 = 0.4658; ...

[0088] It should be noted that, in the above, A, B, C, and D are four differently labeled data files. After sorting each labeled data file according to the size of the weight reference amount, they correspond to the A, B, C, and D labeled data files in order from small to large. In this example, only four labeled data files are used as examples, and the actual number of labels is consistent with the number of data files storing the data.

[0089] The storage switching module calculates the storage weight of each marked data file in the storage data through the hierarchical analysis method, and takes the average storage weight of all marked data files in the storage data as the split threshold; the marked data files in the storage data whose storage weight exceeds the split threshold are screened out and stored using the non-structural storage method.

[0090] It should be noted that structured storage is a method of uniform storage according to organizational rules, and unstructured storage is a method of separate storage that does not follow organizational rules. Separately storing data with excessive storage weight is beneficial to improving storage efficiency and storage security.

[0091] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0092] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0093] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0094] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0095] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A server-based data management system, characterized in that: It includes data acquisition module, structure classification module, splitting judgment module and storage switching module, and the signal connection between each module; The data collection module is used to collect the keyword prefixes and suffixes of the file names and the transmission time interval of the data files in the storage data and pass them to the structure classification module. After receiving the classification results passed in by the structure classification module, the standard processing algorithm of all data files in the storage data and the joint call frequency are detected and passed to the split judgment module; The structural classification module classifies the stored data according to the keyword prefix and suffix of the file name of the data file in the stored data and the transmission time interval, and transmits the classification result back to the data acquisition module; The split judgment module analyzes the maximum processing ratio of the same standard processing algorithm according to the standard processing algorithm of all data files in the storage data, comprehensively considers the maximum processing ratio of the storage data and the joint call frequency to determine whether to pass the storage data into the split storage mechanism, and transmits the judgment result to the storage switching module; The storage switching module selects the structural storage method for storage or data splitting according to the judgment result. Before data splitting, each data file in the storage data is marked, the file update frequency and storage space occupied by each marked data file are obtained, the marked data files are screened, and the screened marked data files are stored in the non-structural storage method; After receiving the standard processing algorithm of all data files in the storage data, the split judgment module calculates and obtains the maximum processing ratio of the storage data. The process is as follows: The split judgment module sorts all data files in the storage data, starting from the standard processing algorithm of the first data file in the storage data, records its category and assigns a value of 1, and after the assignment is completed, traverses the subsequent data files one by one according to the sorting order. When the standard processing algorithm of the subsequent data file is the same as the standard processing algorithm of the assigned data file, the assignment on the corresponding category is increased by 1. When the standard processing algorithm of the subsequent data file is different from the standard processing algorithm of the assigned data file, its category is recorded and assigned a value of 1 until all data files in the storage data are traversed; The category with the largest assigned value is selected from the stored data and recorded as the largest category. The ratio of the assigned value of the largest category to the total assigned values ​​of the categories of all data files in the stored data is taken as the maximum processing ratio of the stored data.

2. A server-based data management system according to claim 1, characterized in that: The data acquisition module obtains the keyword prefixes and suffixes of the file names of each data file in the stored data, counts the types of prefixes and suffixes, obtains the number of keyword prefixes of various types of file names and the number of keyword suffixes of various types of file names, and merges them into a prefix and suffix data set, uses the maximum value selected from the prefix and suffix data set to calculate the ratio with the total number of all data files in the stored data, and uses the calculation result as the prefix and suffix duplication percentage; The average of the transmission time intervals of different data files in the stored data is taken as the mean transmission time. The data acquisition module passes the prefix and suffix duplication percentage of the stored data and the mean transmission time to the structure classification module.

3. A server-based data management system according to claim 2, characterized in that: After receiving the prefix and suffix duplication percentage and the mean transmission time of the stored data, the structure classification module uses the geometric mean method to classify the stored data into high structure data or low structure data. The specific steps are as follows: Data preprocessing: The mean transmission time of the stored data is normalized using logarithmic changes, using the logarithmic change formula: y=ln x, where x is the mean transmission time of the stored data, and y is the normalized result of the mean transmission time of the stored data; Geometric mean analysis: The geometric mean formula is constructed by normalizing the percentage of prefix and suffix duplication of the stored data with the mean of the transmission time: , where y is the normalized result of the mean value of the transmission time of the stored data, z is the percentage of prefix and suffix repetition of the stored data, and L is the result of the geometric mean formula, which is used as the data classification coefficient; Data classification: When the data classification coefficient exceeds the preset classification threshold, the stored data will be classified as high-structure data and the structural storage method will be selected for storage; when the data classification coefficient is lower than the classification threshold, the stored data will be classified as low-structure data and transferred to the data acquisition module.

4. A server-based data management system according to claim 3, characterized in that: After receiving the classification results transmitted by the structure classification module, the data acquisition module accesses the log records to detect the standard processing algorithms of all data files in the stored data and the joint call frequency between data; The data acquisition module selects a period of time as the analysis time, analyzes each data file in the storage data within the analysis time, counts the ratio of the number of joint calls of each data file in the storage data to the total number of calls as the call frequency of the corresponding data file, and selects the median from the call frequencies of all data files as the joint call frequency of the storage data; The joint call is to call other data files on the basis of calling the current data file in the storage data within a preset time interval.

5. A server-based data management system according to claim 1, characterized in that: The split judgment module uses the polynomial regression method to analyze the stored data based on the maximum processing ratio of the stored data and the joint call frequency, and determines whether to import the stored data into the split storage mechanism. The specific process is as follows: The maximum processing ratio of stored data and the joint call frequency are used to construct a polynomial regression formula to calculate the regression results: , where a is the maximum processing ratio of stored data, b is the frequency of joint calls of stored data, and c is the reconciliation parameter. and are the parameter weights of a and b respectively, and D is the regression result; The regression result obtained by calculating the maximum processing ratio of the stored data and the joint call frequency is compared with the preset judgment threshold. When the regression result exceeds the judgment threshold, the judgment result is set to 0; otherwise, the data result is set to 1.

6. A server-based data management system according to claim 5, characterized in that: The storage switching module processes the judgment result transmitted by the split judgment module as follows: When the judgment result is 0, the data is not split and the stored data is stored using the structured storage method; When the judgment structure is 1, the data is split and each data file in the stored data is marked differently; The storage switching module selects a period of time as the update time, records the number of file updates of each marked data file within the update time, uses the number of file updates of the marked data file as the file update frequency of the corresponding marked data file, records the occupied space after each update of the marked data file within the update time and calculates the average value as the occupied space of the corresponding marked data file.

7. A server-based data management system according to claim 6, characterized in that: The storage switching module comprehensively stores the file update frequency and storage space of each marked data file in the data using the hierarchical analysis method to set the storage weight of each marked data file, and screens the marked data files according to the storage weight. The specific steps are as follows: Preliminary data processing: standardize the space occupied by each labeled data file; Calculate the weight reference: sum the file update frequency of each marked data file and the occupied space after normalization, and use the sum result as the weight reference; Select a point scale: Select a point scale from 1 to 4, where 1 indicates an equal importance difference, 2 indicates a slight importance difference, 3 indicates a moderate importance difference, and 4 indicates a severe importance difference; Construct feature matrix: construct feature matrix by setting point quantity according to weight reference quantity of each labeled data file; Simplify the feature matrix: normalize each column of the feature matrix so that the sum of each column is 1; Calculate feature vector: Calculate the feature vector of each labeled data file as the storage weight.

8. A server-based data management system according to claim 7, characterized in that: The storage switching module uses the average storage weight of all the marked data files in the storage data as the split threshold; the marked data files in the storage data whose storage weight exceeds the split threshold are screened out and the non-structural storage method is selected for storage.

Images