Performance optimization method, device and equipment for multi-mode query of data vault and medium

Abstract

The invention discloses a performance optimization method, device and equipment for multimode query of a data vault and a medium. The method comprises the following steps: low-density screening of levels: analyzing data storage, and constructing a low-density index according to data features; the data statistics level is used for calculating and counting data source information; block scanning level: evaluating the data size, configuring block scanning parameters according to the data size, and determining the degree of parallelism of parallel scanning; a condition operator push-down level: identifying a push-down query condition, and pushing down the push-down query condition to the data source scanning operator; a cost model calculation level: planning to generate data sorting for a sorting operator through cost model calculation; a memory model calculation level: allocating a memory model according to the size of the memory; according to the cost of each layer, evaluating the performance improvement proportion or multiple brought by optimization of each layer layer by layer; and planning a data processing mode according to a layer-by-layer evaluation result. According to the method, the multi-layer optimization model is adopted, so that the performance of multi-mode query is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular, to a method, device, equipment and medium for optimizing the performance of multi-modal queries in a data vault. Background Art

[0002] When conducting data governance, the data vault serves as a storage and computing engine to provide queries and calculations for the business. With the development of big data, the sources of data have become rich, and different data may exist in different types of data sources. When performing queries or calculations, performance optimization has become complicated.

[0003] Currently, there are three types of data faced by the data vault: structured data, semi-structured data, and unstructured data. Different data exists in different databases. When it comes to queries and calculations involving different data types, it is necessary to achieve efficient data query and calculation among multiple heterogeneous data sources.

[0004] In the query optimization methods in the related art, the data of the data source is usually scanned, evaluated according to the data size and the SQL type, the data distribution method is adjusted, a production plan is generated, table joins and filters are performed, and finally the result is obtained. In practical applications, when it comes to cross-type and cross-data source data queries, the data vault often adopts an inefficient full-table scan method, resulting in a large number of I / O operations, seriously occupying the server memory and CPU resources, and thus affecting the query calculation ability and overall performance of the data vault. Summary of the Invention

[0005] The present invention provides a method, device, equipment and medium for optimizing the performance of multi-modal queries in a data vault, which solves the problem of optimizing the performance of data query and calculation in a multi-modal scenario.

[0006] To achieve the above object, the present application adopts the following technical solutions: In a first aspect, a method for optimizing the performance of multi-modal queries in a data vault is provided, including: Low-density screening layer: Analyze the data storage, construct a low-density index according to the data characteristics, and perform low-density screening based on the low-density index; Data statistics layer: Calculate and statistically analyze the data source information based on the results of the low-density screening; Block scan layer: Evaluate the data size according to the data source information, configure block scan parameters according to the data size, and determine the parallelism of parallel scans; Condition operator pushdown layer: Identify the query conditions that can be pushed down, and push down the query conditions that can be pushed down to the data source scan operator; Cost model calculation layer: Through cost model calculation, generate data sorting for the sorting operator; Memory model calculation layer: Allocate a memory model according to the memory size; Based on the cost of each layer M , layer by layer, evaluate the performance improvement ratio or multiple brought by each layer of optimization; Plan the data processing method according to the results of layer-by-layer evaluation.

[0007] In the first possible implementation manner of the first aspect, the cost of each layer M is calculated by the following formula: M =N*C*k*h*s*w*c, where: N is a constant coefficient, C = the cost of each node, k = the field to be distributed, h = the CPU cost of calculating the hash value on 1 field, s = the CPU cost of the row selection vector remover for each row, w = the network cost of sending 1 row to 1 destination, c = the CPU cost of comparing the incoming row with the row on the heap of size N.

[0008] In the second possible implementation manner of the first aspect, the low-density screening layer, the data statistics layer, the block scan layer, the conditional operator pushdown layer, the cost model calculation layer, and the memory model calculation layer are executed sequentially in order.

[0009] In the third possible implementation manner of the first aspect, the data sorting includes at least partially unordered intermediate process data.

[0010] In the second aspect, a performance optimization device for multi-mode query of a data vault is provided, including: A low-density screening layer module, which is used to analyze data storage, construct a low-density index according to data characteristics, and perform low-density screening based on the low-density index; A data statistics layer module, which is used to calculate and statistically analyze data source information based on the results of the low-density screening; A block scan layer module, which is used to evaluate the data size according to the data source information, configure block scan parameters according to the data size, and determine the parallelism of parallel scanning; A conditional operator pushdown layer module, which is used to identify pushable query conditions and push the pushable query conditions down to the data source scan operator; A cost model calculation layer module, which is used to calculate through a cost model and plan to generate data sorting for the sorting operator; A memory model calculation layer module, which is used to allocate a memory model according to the memory size; A performance improvement calculation module, which is used to layer by layer evaluate the performance improvement ratio or multiple brought by each layer of optimization based on the cost of each layer M , layer by layer, evaluate the performance improvement ratio or multiple brought by each layer of optimization; A data processing planning module, configured to plan a data processing method according to the results of layer-by-layer evaluation.

[0011] In the first possible implementation manner of the second aspect, the costs of each layer M are calculated by the following formula: M =N*C*k*h*s*w*c, where: N is a constant coefficient, C = the cost of each node, k = the field to be distributed, h = the CPU cost of calculating the hash value on 1 field, s = the CPU cost of the row selection vector remover per row, w = the network cost of sending 1 row to 1 destination, c = the CPU cost of comparing the incoming row with the row on the heap of size N.

[0012] In the second possible implementation manner of the second aspect, the low-density screening layer, the data statistics layer, the block scanning layer, the conditional operator pushdown layer, the cost model calculation layer, and the memory model calculation layer are executed sequentially in order.

[0013] In the third possible implementation manner of the second aspect, the data sorting includes at least partially unordered intermediate process data.

[0014] In a third aspect, an electronic device is provided, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the computer program is executed by the processor, the steps of the performance optimization method for multimodal query of the data vault as described in the first aspect are implemented.

[0015] In a fourth aspect, a readable storage medium is provided, on which a program or instruction is stored, and when the program or instruction is executed by a processor, the steps of the performance optimization method for multimodal query of the data vault as described in the first aspect are implemented.

[0016] The performance optimization method for multimodal query of the data vault of the present invention has the following advantages: This application optimizes multimodal query from perspectives such as data storage, plan generation, and operator optimization, adopts a multi-layer optimization model, starts from the place closest to data storage, and sequentially obtains performance improvement, thereby improving the performance of multimodal query. Following the means of starting from the slowest part of the server, the greater the optimization benefit can be obtained. The optimization architecture means starting from the lowest level and going up layer by layer can more effectively enhance the speed and results of optimization.

[0017] The device, electronic device, and readable storage medium corresponding to the performance optimization method for multimodal query of the data vault of the present invention can achieve the same technical effects. To avoid repetition, they are not described here again. Description of the Drawings

[0018] Figure 1 A schematic flowchart of a method for optimizing the performance of multimodal queries in a data vault provided by an embodiment of the present application; Figure 2 A schematic flowchart of another method for optimizing the performance of multimodal queries in a data vault provided by an embodiment of the present application; Figure 3 A schematic structural diagram of a device for optimizing the performance of multimodal queries in a data vault provided by an embodiment of the present application; Figure 4 A schematic structural diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0019] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined purpose, the technical solutions in the embodiments of the present application are clearly described. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of them. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0020] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that such terms can be interchanged under appropriate circumstances so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. generally belong to the same category, and the number of objects is not limited. For example, the first object can be one or multiple. In addition, "and / or" in the specification and claims means at least one of the connected objects, and the character " / " generally indicates an "or" relationship between the associated objects before and after.

[0021] In the present application, the description of the method flow in the specification and the steps in the flowchart in the accompanying drawings of the present invention do not necessarily have to be strictly executed according to the step numbers. The execution order of the method steps can be changed. Moreover, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution.

[0022] The following will be a detailed description of the method, device, equipment, and medium for optimizing the performance of multimodal queries in a data vault provided by the embodiments of the present application in combination with the accompanying drawings and preferred embodiments as follows.

[0023] First, a detailed description will be given of the application scenario of the method for optimizing the performance of multimodal queries in a data vault in the embodiments of the present application.

[0024] Please refer to Figure 1 , the embodiments of the present application provide a method for optimizing the performance of multimodal queries in a data vault, asFigure 1 As shown in Figure 1 , the recommendation method of the embodiment of the present application includes: Step S1, low-density screening stage: Analyze the data storage, construct a low-density index according to the data characteristics, and perform low-density screening based on the low-density index.

[0025] Analyze the data storage, construct an index structure suitable for different types of data. Taking semi-structured data commonly used in data governance as an example, construct a low-density index with appropriate fields as index keys, and count the data characteristics, thereby improving the data screening efficiency.

[0026] Step S2, data statistics stage: Calculate and count the data source information based on the results of the low-density screening.

[0027] Scan the information of the data source, including but not limited to the data volume, data distribution, etc., to provide an accurate cost estimate for subsequent plan generation.

[0028] Step S3, block scan stage: Evaluate the data size according to the data source information, configure the block scan parameters according to the data size, and determine the parallelism of parallel scanning.

[0029] Evaluate the data size, automatically calculate a reasonable block scan, and calculate the optimal parallelism of parallel scanning according to the resources to improve the scanning efficiency.

[0030] Step S4, conditional operator pushdown stage: Identify the query conditions that can be pushed down, and push down the query conditions that can be pushed down to the data source scan operator.

[0031] The judgment conditions are generally query conditions that can be pushed down. Pass such query conditions directly to the scan operator, so that it can perform preliminary filtering while reading the data, reducing unnecessary data transmission and processing volume, and improving the data screening efficiency.

[0032] Step S5, cost model calculation stage: Through cost model calculation, for the sorting operator, plan to generate data sorting.

[0033] The data sorting includes at least partially unordered intermediate process data.

[0034] In the specific implementation process, it may not be necessary for the intermediate process data to be ordered, and only the final data needs to be ensured to be ordered. Even if the data is sorted in the middle, subsequent data operations may still cause the data to become unordered. Therefore, through cost model calculation, the plan is generated to reduce the data sorting in the intermediate process.

[0035] For example: The cost model will evaluate the delay of the sorting operator, first perform aggregation filtering operations and compare with sorting first and then performing aggregation filtering and other operations. Through cost evaluation of such different operations, select the one with less cost, thereby reducing the data sorting in the intermediate process.

[0036] Step S6, Memory Model Calculation Level: Allocate a memory model according to the memory size.

[0037] Allocate a reasonable memory model, automatically set the batch block, use server resources more efficiently, and improve the join efficiency.

[0038] Step S7, According to the costs of each level, evaluate the performance improvement ratio or multiple brought by each layer of optimization layer by layer; among them, the costs of each level M Are calculated by the following formula: M =N*C*k*h*s*w*c, In the formula: N is a constant coefficient, C = the cost of each node, k = the fields to be distributed, h = the CPU cost of calculating the hash value on 1 field, s = the CPU cost of the row selection vector remover for each row, w = the network cost of sending 1 row to 1 destination, c = the CPU cost of comparing the incoming row with the row on the heap of size N.

[0039] Specifically, the constant coefficient N represents the impact of the data scale on the overall calculation, and the specific value depends on the data scale, usually the total number of data rows or the shard size of the task. The cost C of each node includes resource consumption such as CPU, memory, and network, C = C_cpu + C_mem + C_net, and needs to be measured according to the hardware configuration. The number of fields k to be distributed is directly equal to the number of fields participating in the distribution, such as the number of grouping keys. The CPU cost h of calculating the hash value on 1 field, h = f(field_length, hash_algorithm), depends on the field length and the hash algorithm. The CPU cost s of the row selection vector remover for each row, s = g(select_vector_complexity), depends on the select vector complexity. The network cost w of sending 1 row to 1 destination, w = h(row_size, net_bandwidth, latency), considers the row size and network latency. The CPU cost c of comparing the incoming row with the row on the heap, c = f(compare_complexity, row_size, heap_size).

[0040] During specific implementation: Determine the hash cost (h) of each field: By measuring or estimating the field length and the hash algorithm complexity; The cost of the row selection vector remover (s): Calculate according to the conditional complexity of the selection vector; Network cost (w): Estimated using the formula w = row_size / bandwidth + latency; Comparison cost (c): Estimated based on the size of the heap and the complexity of the comparison logic; Node cost (C): Decomposed into the sum of CPU, memory, and network costs.

[0041] Step S8, according to the results of layer-by-layer evaluation, plan the data processing method.

[0042] Each layer realizes an optimized step-by-step improvement by effectively reducing the costs of C, k, h, s, w, c, plans a reasonable data processing method for the data, and continues to improve the efficiency of join. It includes: Distributed data distribution: Ensure load balancing according to hash distribution; Grouping and sorting: Optimized using distributed merge sort; Combination of batch processing and streaming processing: Select batch processing or streaming processing mode according to the data scale; Layer-by-layer optimization: Adjust the task allocation strategy to reduce network or CPU costs; Cache and intermediate result storage: Use memory cache to store repeatedly calculated data; Dynamic resource allocation: Dynamically adjust resources according to task complexity; Result merging: Perform final aggregation after local processing at the node is completed.

[0043] See Figure 2 In the above performance optimization method for multi-modal query of the data vault, the low-density screening layer level, data statistics layer level, block scanning layer level, conditional operator push-down layer level, cost model calculation layer level, and memory model calculation layer level are executed sequentially in order. An optimization architecture is formed that starts from the lowest level and goes up layer by layer.

[0044] Based on the above technical solution, this application optimizes multi-modal query from the perspectives of data storage, plan generation, operator optimization, etc., adopts a multi-level optimization model, starts from the place closest to data storage, and sequentially obtains performance improvements, thereby improving the performance of multi-modal query. Following the means of the optimization architecture that starts from the slowest part of the server and can obtain the greatest optimization benefits, starting from the lowest level and going up layer by layer, can more effectively enhance the speed and results of optimization. And to measure the overall improvement degree brought by this multi-level optimization strategy, it is obtained by multiplying the improvement ratios (or called "multiples") obtained at each level one by one. This calculation process is like multiplying a series of numbers called k1, k2 until kn together, and the final "magic total" is the overall optimization multiple we mentioned.

[0045] See Figure 3, corresponding to the embodiment of the performance optimization method for multi-modal query of the above data vault, an embodiment of the present application provides a performance optimization device for multi-modal query of a data vault, and the device includes: A low-density screening layer module 1001, configured to analyze data storage, construct a low-density index according to data characteristics, and perform low-density screening based on the low-density index; A data statistics layer module 1002, configured to calculate and statistically analyze data source information based on the result of the low-density screening; A block scan layer module 1003, configured to evaluate the data size according to the data source information, configure block scan parameters according to the data size, and determine the parallelism of parallel scanning; A conditional operator push-down layer module 1004, configured to identify query conditions that can be pushed down, and push down the query conditions that can be pushed down to a data source scan operator; A cost model calculation layer module 1005, configured to calculate through a cost model and generate data sorting for a sorting operator; A memory model calculation layer module 1006, configured to allocate a memory model according to the memory size; A performance improvement calculation module 1007, configured to calculate according to the costs of each layer M , and layer by layer evaluate the performance improvement ratio or multiple brought by each layer of optimization; A data processing planning module 1008, configured to plan a data processing method according to the results of layer-by-layer evaluation.

[0046] Further, the costs of each layer M are calculated by the following formula: M =N*C*k*h*s*w*c, where: N is a constant coefficient, C = the cost of each node, k = the field to be distributed, h = the CPU cost of calculating the hash value on 1 field, s = the CPU cost of the row selection vector remover for each row, w = the network cost of sending 1 row to 1 destination, c = the CPU cost of comparing the incoming row with the row on the heap of size N.

[0047] Further, the low-density screening layer, the data statistics layer, the block scan layer, the conditional operator push-down layer, the cost model calculation layer, and the memory model calculation layer are executed in sequence.

[0048] Further, the data sorting includes at least partially unordered intermediate process data.

[0049] The above performance optimization device for multimode query of the data repository implements the steps and various processes of the above embodiments of the performance optimization method for multimode query of the data repository, and can achieve the same technical effects. To avoid repetition, they will not be elaborated here.

[0050] See Figure 4 , corresponding to the above embodiments of the performance optimization method for multimode query of the data repository, an embodiment of the present application provides an electronic device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps and various processes of the above embodiments of the performance optimization method for multimode query of the data repository, and can achieve the same technical effects. To avoid repetition, they will not be elaborated here.

[0051] The memory 1009 can be used to store software programs and various data. The memory 1009 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data. Among them, the first storage area can store an operating system, applications or instructions required for at least one function (such as a sound playback function, an image playback function, etc.). In addition, the memory 1009 can include volatile memory or non-volatile memory, or the memory 1009 can include both volatile and non-volatile memory. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDR SDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM), and a direct rambus random access memory (DRRAM). The memory 1009 in the embodiments of the present application includes but is not limited to these and any other suitable types of memory.

[0052] The processor 1010 may include one or more processing units; optionally, the processor 1010 integrates an application processor and a modem processor. Among them, the application processor mainly processes operations related to the operating system, user interface, application programs, etc., and the modem processor mainly processes wireless communication signals, such as a baseband processor. It can be understood that the above-mentioned modem processor may not be integrated into the processor 1010 either.

[0053] Corresponding to the embodiment of the performance optimization method for multimode query of the above data vault, the embodiment of the present application also provides a readable storage medium. A program or instruction is stored on the readable storage medium. When the program or instruction is executed by a processor, the steps and processes of the embodiment of the performance optimization method for multimode query of the above data vault are implemented, and the same technical effects can be achieved. To avoid repetition, it will not be elaborated here.

[0054] Among them, the processor is the processor in the electronic device described in the embodiment of the present application. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk, or an optical disc, etc.

[0055] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including that element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in a reverse order according to the functions involved. For example, the methods described may be performed in an order different from that described, and various steps may be added, omitted, or combined. Additionally, the features described with reference to certain examples may be combined in other examples.

[0056] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described example methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to enable a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present application.

[0057] It can be understood that the embodiments of the present application have been described above in conjunction with the accompanying drawings. However, the present application is not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative and not restrictive. As those skilled in the art know, without departing from the spirit and scope of the present invention, these features and embodiments can be variously changed or equivalently replaced. In addition, those of ordinary skill in the art can modify these features and embodiments under the inspiration or teaching of the present application to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of the present application belong to the scope protected by the present invention.

Claims

1. A performance optimization method for multi-mode query of a database, characterized in that: include: Low-density screening layer: Analyze data storage, build low-density indexes according to data characteristics, and perform low-density screening based on low-density indexes; Data statistics level: based on the results of the low-density screening, calculate and store data source information; Block scanning level: evaluating the data size according to the data source information, configuring block scanning parameters according to the data size, and determining the parallelism of parallel scanning; Condition operator push-down level: identifying query conditions that can be pushed down, and pushing the query conditions that can be pushed down to the data source scanning operator; Cost model calculation level: Through cost model calculation, the data sorting is planned for the sorting operator; Memory model calculation level: allocate memory model according to memory size; According to the cost M of each layer, evaluate the performance improvement ratio or multiple brought by the optimization of each layer layer by layer; Plan data processing methods based on the results of layer-by-layer assessment.

2. The performance optimization method for multi-mode query of a database according to claim 1 is characterized in that: The cost M of each level is calculated by the following formula: M=N*C*k*h*s*w*c, Where: N is a constant coefficient, C = cost per node, k = field to distribute, h = CPU cost of computing hash value on 1 field, s = CPU cost of selecting vector remover per row, w = network cost of sending 1 row to 1 destination, c = CPU cost of comparing incoming row with rows on heap of size N.

3. The performance optimization method for multi-mode query of a database according to claim 1 is characterized in that: The low-density screening level, data statistics level, block scanning level, conditional operator push-down level, cost model calculation level and memory model calculation level are executed in sequence.

4. The performance optimization method for multi-mode query of a database according to claim 1 is characterized in that: The data sorting includes at least partially unordered intermediate process data.

5. A performance optimization device for multi-mode query of a database, characterized in that: include: The low-density screening hierarchical module is used to analyze data storage, build low-density indexes according to data characteristics, and perform low-density screening based on low-density indexes; A data statistics hierarchy module, used to calculate and store data source information based on the results of the low-density screening; a block scanning hierarchy module, for evaluating the data size according to the data source information, configuring block scanning parameters according to the data size, and determining the parallelism of parallel scanning; A condition operator push-down hierarchy module, used to identify query conditions that can be pushed down, and push the query conditions that can be pushed down to the data source scanning operator; The cost model calculation hierarchy module is used to calculate the cost model and plan the data sorting for the sorting operator; The memory model calculation hierarchy module is used to allocate memory models according to the memory size; The performance improvement calculation module is used to evaluate the performance improvement ratio or multiple brought by the optimization of each layer layer by layer according to the cost M of each layer; The data processing planning module is used to plan the data processing method based on the results of layer-by-layer evaluation.

6. The performance optimization device for multi-mode query of a database according to claim 5, characterized in that: The cost M of each level is calculated by the following formula: M=N*C*k*h*s*w*c, Where: N is a constant coefficient, C = cost per node, k = field to distribute, h = CPU cost of computing hash value on 1 field, s = CPU cost of selecting vector remover per row, w = network cost of sending 1 row to 1 destination, c = CPU cost of comparing incoming row with rows on heap of size N.

7. The performance optimization device for multi-mode query of a database according to claim 5, characterized in that: The low-density screening level, data statistics level, block scanning level, conditional operator push-down level, cost model calculation level and memory model calculation level are executed in sequence.

8. The performance optimization device for multi-mode query of a database according to claim 5, characterized in that: The data sorting includes at least partially unordered intermediate process data.

9. An electronic device, characterized in that: The electronic device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the performance optimization method for multi-modal query of a database vault as claimed in any one of claims 1 to 4.

10. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of the performance optimization method for multi-mode query of the database vault as described in any one of claims 1 to 4 are implemented.

Images