KR101311409B1 - Partition scan method and device, memory system, and data alignment method using partial index rid alignment - Google Patents

Abstract

The present invention scans a partition through a partial index Low IDentifier (RID) sorting according to a key value in a memory including a database table storing at least one record and an index entry in which the index of the record is collected. A method, comprising: sorting an index entry according to the key value by a memory controller; Dividing, by the memory controller, the index entries arranged according to the key value by the partition unit; And reading, by the memory controller, the records indicated by the index RIDs from the database table in the index RID order of the partitions, and performing an internal sort. A partition scan method and apparatus, a memory system, and a data sorting method are provided.

Description

PARTITION SCAN METHOD AND DEVICE, MEMORY SYSTEM, AND DATA ALIGNMENT METHOD USING PARTIAL INDEX RID ALIGNMENT}

The present invention relates to a partition scan method and apparatus, a memory system, and a data sorting method. More particularly, the present invention relates to a partition scan method for reducing the number of disk I / Os that require the most time during the sorting process by performing a single read and write operation. A device, a memory system including the device, and a data alignment method.

In general, small wireless terminals such as mobile phones and PDAs are equipped with an embedded DBMS (Database Management System) for efficient database management. These DBMSs manage data using RTOS (Realtime OS) or middleware file system and memory functions. It provides system software that can directly handle tables, and SQL api that can manage data using Structured Query Language (SQL), a 3G language.

1 is a diagram showing a conventional DBMS (DBMS) structure. As shown in FIG. 1, a general DBMS structure includes a flash memory 150 in which a data file is stored, and a RAM 100 for processing data. The flash memory 150 stores the database table 160 including the table record 161, the index entry 170, the active log 180, and the like, and the RAM 100 arranges the database table 160. Memory space for sort (120), index buffer 121 for processing the index entry 170, log buffer 122 for processing the log file 180, buffer for processing the database table 160 The storage manager 110 for controlling the management operation of the cache 123, the database table and the index entry 170, and a recovery manager for recovering damaged data during data input and output.

The database table 160 is data having a concept of rows and columns, for example, data stored in an employee number, name, age, and department in a table called employee. The database table 160 may include a plurality of records 161. The records 161 may be a combination of various fields to systematically manage data.

The database table 160 may include data such as an address book, a memo for schedule management, an anniversary, schedule information, MP3 data, media, a call log, a short message, and the like, and an index entry 170. In the table 160, index information (RID: Row IDentifier) of various data stored in the database table 160 and key values for distinguishing data are stored.

In addition, the index information RID stored in the index entry 170 is formed in a tree structure. In the tree structure, the index information RID is stored in one upper tree node, and two lower children are located below the upper tree node. Tree nodes have a linked structure.

RID simply says the location of the row. In other words, you can think of it as a location pointer that points to the location of the data because you need to know where the stored row is to read the data. In fact, when we use the select statement, we need to know this RID and know how many rows are in the block such as datafile # and block # that the RID points to, so we can know the exact row data and give the value to the user.

In the conventional data scanning method of the tree structure, when a key value for searching is given, the given key value and the index key value existing in the tree node are compared one by one. If a matching index key value is found, the data is read from the database table using the index information and the information is stored in the corresponding location or read and output. Conventional scanning method has to read and compare the data connected to all index information (RID) in the tree to find the index information (RID) corresponding to a given key value when searching for specific data, so that the data can be read buffer When there is not enough space, the frequency of accessing the data increases, which causes a problem of not being able to search quickly.

In particular, in an environment in which a buffer cache is maintained and used for a data area such as a database of a small wireless terminal, when the frequency of data access is increased, there is a problem in that performance is adversely affected.

In addition, the index-based approach using the conventional index RID sorting has a problem of generating additional sorting operations because the resulting data is sorted in RID order rather than a key value input by the user.

Korean Unexamined Patent Publication No. 1020070007264 ("Non-volatile Memory and Method with Non-Sequential Update Block Management", HDD Disc Corporation, Jan. 15, 2007) Korean Registered Patent 100981317 ("Method for providing grouped and classified tree structure indexes of DBS for small wireless terminals and information retrieval method using them", Innerbit Co., 2010.09.10 published)

An object of the present invention is to write to the disk more than once if there is not enough memory when performing the sort, in order to prevent this by performing a sort with one read and write the most time during the sorting process The present invention provides a method and a device for scanning a partition through partial index RID sorting, a memory system including the same, and a data sorting method for improving performance by reducing the number of disk I / Os.

In addition, another object of the present invention is a partition scan method and apparatus using a partial index RID sort that does not cause additional sorting operation because the resulting data is sorted by the user's key value rather than the RID order and memory including the same To provide a system.

According to an embodiment of the present invention, a partition scan method using a partial index RID sorting is performed according to a key value for a memory including a database table storing at least one record and an index entry collecting the indexes of the records. CLAIMS What is claimed is: 1. A method of scanning a partition through an index Low IDentifier (RID) alignment, the method comprising: sorting an index entry according to the key value by a memory controller; Dividing, by the memory controller, the index entries arranged according to the key value by the partition unit; And reading, by the memory controller, a record indicated by the index RID from a database table in the index RID order of the partitions, and performing an internal sort.

In the dividing into the partition unit, the memory controller may calculate the size and number of partitions and divide the index entry arranged in the partition unit according to the key value according to the calculated size and number of partitions.

In the partitioning step, the size and number of partitions may be calculated by comparing the sort size of the memory allocated to the sort to the size of the records in the table.

In the partitioning step, the size of the partition may be the same as the memory allocated to the alignment.

In the partitioning step, the number of partitions may be proportional to the size of a record in the table.

According to an embodiment of the present invention, a partition scanning apparatus through partial index RID sorting may be performed according to a key value with respect to a memory including a database table storing at least one record and an index entry collecting the indexes of the records. An apparatus for scanning a partition through an index Low IDentifier (RID) sorting, comprising: a key value sorter for sorting an index entry according to the key value; A partition dividing unit dividing the index entries sorted according to the key value into the partition units; And a partition scan unit that reads the records indicated by the index RIDs from a database table in the partitioned partitions in the index RID order and performs an internal sort.

The partition dividing unit calculates the size and number of partitions and divides the index entries arranged according to the key value according to the calculated size and number of partitions in the partition unit.

The partition partitioner may calculate the size and number of partitions by comparing the size of the memory allocated to the sort with the size of the records in the table.

The partition partitioner may calculate the size of the partition to be equal to the memory allocated to the alignment.

The partition dividing unit may calculate the number of partitions to be proportional to the size of a record in the table.

In the memory system through the partial index RID sorting according to an embodiment of the present invention, a memory system capable of scanning a partition through the partial index row IDentifier (RID) sorting according to a key value, A memory including an index entry storing an index of the record and a database table storing at least one record, and sorting the index entry according to the key value, and sorting the index entry according to the key value in the partition unit. After dividing into, the memory controller may perform an internal sorting of the partitions in the index RID order.

The memory controller includes a key value sorter for sorting the index entry according to the key value; A partition dividing unit for calculating the size and number of the partitions and dividing the index entries arranged according to the key values according to the calculated size and number of partitions in the partition unit; And a partition scan unit that reads the partitioned partitions from the table in the index RID order and performs an internal sort.

The memory controller may calculate the size and number of partitions by comparing the size of the memory allocated to the alignment with the size of the records in the table.

The memory controller may calculate the size of the partition equal to the memory allocated to the alignment.

The memory controller may calculate the number of partitions in proportion to the size of a record in the table.

The memory may be either a flash memory or a hard disk.

According to an embodiment of the present invention, a method of sorting data through partial index RID sorting may include a key value for a memory including a database table storing at least one record and an index entry in which indexes of the records are collected. A method for sorting data through partial index row IDentifier (RID) sorting, the method comprising: sorting an index entry according to the key value by a memory controller; Dividing, by the memory controller, the index entries sorted according to the key value by the partition unit; And performing, by the memory controller, internal sorting of the divided partitions according to the key value.

In the partitioning step, the memory controller may calculate the size and number of partitions and divide the index entries arranged according to the key value according to the calculated size and number of partitions into the partition units.

In the partitioning step, the size and number of partitions may be calculated by comparing the sort size of the memory allocated to the sort to the size of the records in the table.

In the dividing step by the partition unit, the size of the partition may be the same as the memory allocated to the alignment.

In the partitioning step, the number of partitions may be proportional to the size of records in the table.

According to the method and apparatus for partition scanning through partial index RID sorting, a memory system including the same, and a data sorting method of the present invention, a flash memory is designed because it is designed to minimize random access and efficiently process sequential access. In the enterprise storage means used, the efficiency can be greatly increased.

In addition, according to the partition scanning method and apparatus of the present invention, a memory system including the same, and a data sorting method, the desired data can be quickly found in a desired order in a large-capacity server. Not only can it benefit from existing hard disks, it can also contribute to developing flash-related software programs.

1 is a diagram showing a conventional DBMS (DBMS) structure,
2 is a diagram schematically illustrating a partition scanning method through partial index RID alignment according to an embodiment of the present invention;
3 is a flowchart illustrating a partition scan method using partial index RID alignment according to an embodiment of the present invention;
4 is a detailed flowchart illustrating a key value sorting step of a partition scan method using partial index RID sorting according to an embodiment of the present invention;
5 is a detailed flowchart illustrating a partitioning step in a partition scanning method through partial index RID alignment according to an embodiment of the present invention;
6 is a detailed flowchart of a partition scanning step of a partition scanning method through partial index RID alignment according to an embodiment of the present invention;
7 is a block diagram schematically illustrating a partition scan apparatus using partial index RID alignment according to an embodiment of the present invention;
8 is a block diagram schematically illustrating a memory system with partial index RID alignment according to an embodiment of the present invention;
9 is a flowchart schematically illustrating a data sorting method through partial index RID sorting according to another embodiment of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail.

It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may be present in between. Should be. On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the relevant art and are to be interpreted in an ideal or overly formal sense unless explicitly defined in the present application Do not.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In order to facilitate the understanding of the present invention, the same reference numerals are used for the same constituent elements in the drawings and redundant explanations for the same constituent elements are omitted.

2 is a diagram schematically illustrating a partition scan method through partial index RID alignment according to an embodiment of the present invention. As shown in FIG. 2, the index entry 210 is storage in which index information of data to be searched by a user is stored. When the index entry 210 is sorted according to a key value 230 corresponding to a user's desired condition clause and the data is scanned in partition units, the index entry 210 is first sorted according to the key value 230. To be done. In order to do so, the key value 230 needs to be clarified, and thus, a process of searching for the key value 230 that satisfies the condition within the index entry 210 is required.

Conditional clause means the condition that user wants to search. When a user inputs a desired condition clause, the index entry 210 searches for a key value 230 corresponding to the condition clause. For example, if data related to integers 1 to 50 are entered as conditional clauses, the key value 230 may be searched for an integer of 1, 2, ..., 50 from the index entry 210.

Then, the data entry 210 is divided into partitions according to the searched key value 230. Specifically, in order to divide the data entry 210 into partitions according to the key value 230, a process of calculating the size and number of partitions is required. When the size and number of partitions are calculated, the index entry 210 is divided into partition units according to the calculated size and number of partitions. Referring to FIG. 2, it can be seen that the index entry 210 is arranged in the order of the key value 230 from partition 1 250-1 and partition 2 250-2 to partition N 250 -N. . After that, the partitioned partitions are fetched from the database table 270 by indexing the table records indicated by the index RIDs, and each partition is internally sorted by key values. Since each partition internally sorted by key values is partitioned in the order of key values 230, the partitions can be immediately output without a separate merging process between the partitions. As shown in FIG. 2, output 1 290-1 indicates that partition 1 250-1 is patched in RID order and sorted again by key values. Output 2 290-2 shows partition 2 ( It is patched in RID order of 250-2) and sorted by key value again, and output N (290-N) is patched in RID order of partition N (250-N) and then sorted by key value again.

3 is a flowchart schematically illustrating a partition scan method using partial index RID alignment according to an embodiment of the present invention. As shown in FIG. 3, the partition scan method using the partial index RID sorting according to the present invention provides a method for scanning a partition through the partial index RID sorting according to a key value corresponding to a conditional clause. Sorting the index entries according to the key value corresponding to the step 300, the memory controller partitioning the index entries sorted according to the key value in a partition unit 310, and the memory controller indexes the index RIDs for the partitions. And sequentially reading the table records indicated by the index RIDs from the database table and performing an internal sort. Description of each step will be described below with reference to detailed flowcharts of each step of FIGS. 4 to 6.

4 is a detailed flowchart of a key value sorting step 300 of a partition scan method using partial index RID sorting according to an embodiment of the present invention. As shown in FIG. 4, the partition scan method according to the present invention first searches for a key value corresponding to a conditional clause in an index entry (400). As described above, since the user inputs a condition clause to be searched, it is necessary to find a key value corresponding to the condition clause desired by the user in the index entry. Once the key value is found, the index entries are sorted according to the key value (410).

5 is a detailed flowchart of a partition partitioning step 310 of a partition scanning method through partial index RID alignment according to an embodiment of the present invention. As shown in FIG. 5, the partitioning step first involves comparing the size of the memory allocated to the sort with the size of the table record (500). If the amount of memory allocated for the sort is larger than the size of the entire table record, index RID sorting can be done without dividing the index entries into partitions. Large database tables have huge amounts of table records, so this would be the exception.

If the size of the entire table record is larger than the memory allocated for the sort, the index entries cannot be read from the memory allocated for the sort, so that the index entries are large enough to read from the memory allocated for the sort. It must be divided. In dividing data into partitions, the size and number of partitions play an important role. If the size of the partition is too small, the number of entries and exits in the memory can be large and inefficient. In particular, since the memory entry / exit time is very slow compared to the main memory access time, minimizing the number of entry / exit may be the most important performance improvement method. Conversely, if the size of the partition is too large, you may not be able to see the intended performance improvement because you may be out of the memory allocated for sorting.

Therefore, it is important to ensure that partitions have an appropriate size and number. For this purpose, the size of each partition is calculated by comparing the size of the memory allocated to the sort with the size of the table record and the range of records that can be internally sorted. (510). In other words, the size of each partition is calculated to be equal to the size of memory allocated for sorting. The number of partitions is calculated in proportion to the amount of table records (520). Therefore, the number of partitions can be calculated by dividing the size of the table record by the size of the partition. For example, if you have 100 kbytes of table record data and the memory allocated for the sort has a size of 10 kbytes, the partition size will be 10 kbytes equal to the memory allocated for the sort, and the number of partitions will be 100 ÷ 10. It becomes ten. After the calculation of the number and size of partitions is completed, the step of dividing the index entry into partition units according to the calculated size and number of partitions is performed (530).

6 is a detailed flowchart of a partition scan step 320 of a partition scan method through partial index RID alignment according to an embodiment of the present invention. As shown in FIG. 6, the data of the partitioned partition unit reads the table record indicated by the RID from the database table through index RID sorting (600). Thereafter, the final result of performing internal alignment on the partition from which the record is read from the table is output (610). It is preferable to repeat the above process until the end of the table.

Through the above approach, the relationship between each partition is sorted by key value first, and the internal sorting is performed only for the inside of the partition to complete the sort. You can complete the alignment. In addition, regardless of the total amount of data and the amount of memory allocated for sorting, the final result can be calculated at the cost of reading and writing data once, minimizing the most expensive disk accesses during database operations, thereby helping to improve performance. It is.

The scan cost of a partition when the partition is scanned using the partition scan method as described above can be found through Equation 1.

Here, the blevel (Branch Level) value means the depth of the index tree, and selectivity means the selectivity. For example, if there are condition clauses a, b, c, d, and e and the key value a is selected, the selectivity of a becomes one fifth, that is, 20%. leaf_block represents the lowest level number of blocks in the tree structure. N in the middle means the number of partitions, and block_utilization_factor means the number of blocks that must be actually read in order to read the table in the range.

To find the block_utilization_factor, first find the amount of memory allocated for the sort, and then find the average length of the records in the table. It then finds the number of records that can be sorted internally, and sorts the page number of the index entry by the number of memory allocated for sorting. Then, you can get the value of block_utilization_factor by finding the number of records in the table that are not duplicated.

7 is a block diagram schematically illustrating a partition scan apparatus through partial index RID alignment according to an embodiment of the present invention. As shown in FIG. 7, the partition scan apparatus using the partial index RID alignment according to an embodiment of the present invention includes a key value sorter 700, a partition divider 710, and a partition scan unit 720. Can be done.

The key value sorter 700 searches for the key value corresponding to the conditional clause in the index entry 701 and sorts the index entry 701 according to the searched key value.

The partition dividing unit 720 compares the size of the memory allocated to the sort with the size of the table record to calculate the size and number of partitions that can appropriately divide the data as described above, and accordingly the index entry 701 Split into partitions. It is desirable to calculate the size and number of partitions to a size that allows internal sorting. Therefore, the size of each partition is calculated as the size of the memory allocated for sorting, and the number of partitions is calculated in proportion to the amount of table records, which is the total data divided by the size of the partition. The partition dividing unit 710 divides data into partition units according to the calculated size and number of partitions. When partitioning the data in the partition partitioning unit 710, the relations between partitions are already sorted in the order of key values, so that only the intra-partition sorting can complete the sorting of the entire data without a separate data merging process.

The partition scan unit 720 reads records in the partition unit data divided as described above from the database table 702 in the index RID order. The partition from which records are read from the database table 702 performs internal sorting, and the final result may be output through the result output unit 730. It is preferable to repeat the above process until the end of the table.

8 is a schematic block diagram of a memory system 800 with partial index RID alignment in accordance with one embodiment of the present invention. Referring to FIG. 7, the key value alignment unit 811, the partition divider 812, and the partition scan unit 813 of FIG. 8 perform the same role as described with reference to FIG. 7. Since the three components 811, 812, and 813 perform a function of controlling the memory 820, the memory controller 810 is configured. The memory controller 810 may control the index entry 821 and the database table 822 in the memory 820 to appropriately partition the partitions and to align the data according to a user's desired condition. Here, the memory 820 includes an index entry 821 that stores index information and a large database table 822. The memory 820 may be a flash memory or a hard disk.

9 is a flowchart schematically illustrating a data sorting method through partial index RID sorting according to another embodiment of the present invention. As shown in FIG. 9, the method for sorting data through partial index RID sorting according to the present invention includes searching for a key value corresponding to a conditional clause in an index and sorting data according to the key value (900). The partitioned index entries may be divided into partition units 910 and partitioned partitions may be internally sorted according to key values 920.

In the data sorting method of the present invention, first, a key value corresponding to a conditional clause is searched in an index. The data is then sorted according to the searched key values. Since the output of the final sort must be output in order of the key values, the data must first be sorted according to the key values.

Then, compare the size of the table records with the memory allocated for the sort. If the amount of memory allocated for the sort is larger than the size of the table record, you do not need to divide the data into partitions. If the size of the table record is larger than the memory allocated to the sort, the table record data cannot be read from memory, so it is divided into enough to read from memory.

Since it is important to partition the partition to have an appropriate size and number, the present invention calculates the size and number of partitions as follows. The size of a partition is calculated by comparing the size of the memory allocated for the sort with the size of the table records, and calculating each partition as the range of records that can be sorted internally. The size of each partition is calculated to be equal to the size of memory allocated for sorting, and the number of partitions is calculated in proportion to the amount of table records. Therefore, the number of partitions can be calculated by dividing the size of the table record by the size of the partition.

After the calculation of the number and size of partitions is completed, the data is sorted by partitioning the data according to the calculated size and number of partitions and internally sorting the data of the partitioned partition units by key values again.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the inventions as defined by the following claims It will be understood that various modifications and changes may be made thereto without departing from the spirit and scope of the invention.

100: RAM 110: Storage Manager
120: memory allocated for sort 121: index buffer
122: log buffer 123: buffer cache
150: flash memory 160: table records
161: Database Table 170: Index Entries
180: active log
210: Index entry 230: Key value
250: Partition 270: Database Tables
290: Output 300: Sort Key Values Step
310: partition partition step 320: partition scan step
400: key value search step 410: key value sort step
500: Memory and Table Record Size Comparison Steps
510: Partition size calculation step 520: Partition number calculation step
530: Partition partition step 600: Record scan step
610: internal sorting step 700: key value sorting unit
701: Index Entry 702: Database Table
710: partition partition unit 720: partition scan unit
730: result output unit 800: memory system
810: memory controller 811: key value sorter
812: partition partition unit 813: partition scan unit
820: memory 821: index entry
822: Database Table 900: Sorting Key Values
910: partition partition step 920: internal sort step

Claims (21)

A method for scanning a partition through a partial index row identifier sorting according to a key value for a memory including an index entry that collects an index of the record and a database table storing at least one record. In
A memory controller sorting index entries according to the key value;
Dividing, by the memory controller, the index entries arranged according to the key value by the partition unit; And
And the memory controller reading the record indicated by the index RID from the database table in the index RID order of the partitions and performing an internal sort on the partitions. Way.
The method of claim 1, wherein the dividing into partition units comprises:
The memory controller calculates the size and number of partitions and partitions the index entries arranged according to the key values according to the calculated size and number of partitions in the partition unit. Scan method.
The method of claim 1, wherein the dividing into partition units comprises:
And calculating the size and number of the partitions by comparing the sort size of the memory allocated to the sort with the size of the records in the table.
The method of claim 3, wherein the dividing into partition units comprises:
And the size of the partition is the same as the memory allocated to the sort.
The method of claim 3, wherein the dividing into partition units comprises:
And the number of partitions is proportional to the size of a record in the table.
A device that scans a partition through a partial index Low IDentifier (RID) sort according to a key value for a memory including an index entry that collects the index of the record and a database table that stores at least one record. In
A key value sorter for sorting index entries according to the key value;
A partition dividing unit dividing the index entries sorted according to the key value into the partition units; And
And a partition scan unit which reads the records indicated by the index RID from a database table in the order of the index RIDs, and performs an internal sort.
The method of claim 6, wherein the partition partition unit,
Computing the size and number of partitions, and partitions the index entries arranged according to the key value according to the calculated size and number of partitions in the partition unit partition partition scan apparatus using a partial index RID sort.
The method of claim 6, wherein the partition partition unit,
And comparing the size of the memory allocated to the sort with the size of the records in the table to calculate the size and number of the partitions.
The method of claim 8, wherein the partition partition unit,
And calculating the size of the partition to be equal to the memory allocated to the sort.
The method of claim 9, wherein the partition partition unit,
And counting the number of partitions in proportion to the size of the records in the table.
A memory system capable of scanning partitions through partial index row identifier sorting according to key values,
A memory including a database table storing at least one record and an index entry in which the index of the record is collected;
Sorting the index entry according to the key value, dividing the index entry sorted according to the key value in the partition unit, and then, in the database table, the records indicated by the index RID in the index RID order. And a memory controller that reads and performs an internal sort.
The method of claim 11, wherein the memory controller is
A key value sorter for sorting the index entry according to the key value;
A partition dividing unit for calculating the size and number of the partitions and dividing the index entries arranged according to the key values according to the calculated size and number of partitions in the partition unit; And
And a partition scan unit that reads the partitioned partitions from the table in the index RID order and performs an internal sort.
The method of claim 11, wherein the memory controller,
And comparing the size of the memory allocated to the sort with the size of the records in the table to calculate the size and number of the partitions.
The memory controller of claim 13, wherein the memory controller comprises:
And calculating the size of the partition equally to the memory allocated to the sort.
The memory controller of claim 13, wherein the memory controller comprises:
And counting the number of partitions in proportion to the size of the records in the table.
The method of claim 11,
And the memory is one of a flash memory or a hard disk.
A method of sorting data through a partial index row identifier sorting according to a key value for a memory including an index entry storing an index of the record and a database table storing at least one record. In
A memory controller sorting index entries according to the key value;
Dividing, by the memory controller, the index entries arranged according to the key value in units of partitions; And
And a memory controller performing internal sorting on the partitions again according to the key value.
18. The method of claim 17, wherein in the partitioning step,
The memory controller calculates the size and number of partitions and divides the index entries arranged according to the key value according to the calculated size and number of partitions into the partition unit. How to sort.
18. The method of claim 17, wherein in the partitioning step,
And calculating the size and number of the partitions by comparing the sort size of the memory allocated to the sort with the size of the records in the table.
20. The method of claim 19, wherein in the dividing step in the partition unit,
And the size of the partition is the same as the memory allocated to the sort.
20. The method of claim 19, wherein in the dividing step in the partition unit,
And the number of partitions is proportional to the size of records in the table.

Images