CN103823865A - Database primary memory indexing method - Google Patents

Database primary memory indexing method Download PDF

Info

Publication number
CN103823865A
CN103823865A CN201410063636.0A CN201410063636A CN103823865A CN 103823865 A CN103823865 A CN 103823865A CN 201410063636 A CN201410063636 A CN 201410063636A CN 103823865 A CN103823865 A CN 103823865A
Authority
CN
China
Prior art keywords
index
memory
file
key
indexing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201410063636.0A
Other languages
Chinese (zh)
Inventor
秦小麟
王胜
朱广蔚
沈尧
王宁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Aeronautics and Astronautics
Original Assignee
Nanjing University of Aeronautics and Astronautics
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Aeronautics and Astronautics filed Critical Nanjing University of Aeronautics and Astronautics
Priority to CN201410063636.0A priority Critical patent/CN103823865A/en
Publication of CN103823865A publication Critical patent/CN103823865A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/22Indexing; Data structures therefor; Storage structures
    • G06F16/2228Indexing structures

Abstract

The invention provides a database primary memory indexing method, and belongs to the technical field of efficient queries about a computer database. According to the database primary memory indexing method, a traditional file reading and writing mode is replaced with the mode that an index file is imported and exported through memory mapping, so that access efficiency is improved. An index head is established at the starting address of a mapping area, basic index information is stored, and various efficient aggregation operations are supported at the same time. In a mapped memory, index records are read in according to a data table to establish an index tree, wherein the index tree is improved specific to the shortcomings of B-Tree indexing, a repeat key can be supported efficiently, and memory space and the complexity of the tree are reduced. As the number of the index records increases, a file with a preset size will have an overflow risk, and the database primary memory indexing method can be used for effectively expanding the index file and guarantee that the index records can be contained without waste of the memory space.

Description

A kind of database main memory indexing means
technical field
The present invention relates to a kind of database main memory indexing means, belong to the efficient inquiring technology of Computer Database field.
background technology
Along with the development of main memory technology, the capacity of main memory is increasing, and increasing system is based on main memory, and as main storage data base, the traditional database being about to based on disk has been moved in internal memory, has improved significantly system performance.In database, in order to improve search efficiency, often on a certain major key in tables of data, create index, generating indexes file, key assignments is mapped directly to memory address, inquiry is according to key word, inquires after corresponding memory address directly access disk character sensing strip, can improve significantly like this search efficiency.
In traditional indexed mode, often adopt frequently file read-write to operate index file is read in to internal memory, this can expend a large amount of I/O time, can not guarantee to read in the integrality of internal memory index simultaneously, therefore has influence on the stability of search index.In the time that the data in his-and-hers watches are modified, increased and delete, index needs to safeguard, traditional index is frequently accessed disk file and is difficult to safeguard due to needs, has so also just reduced the maintainability of tables of data.
Along with the development of information age, the data volume of database table storage is increasing, operates for so large file, and it will be very low adopting traditional file mode of operation to create index efficiency, and memory-mapped technology can save time effectively.Memory-mapped refers to a File Mapping in internal memory, process can be as access memory access file, and needn't call traditional file read-write function.Due to before mapping, all to set in advance mapped file size, in the time constantly having new record to insert, can cause file to overflow like this, now must index file dilatation not wasted to memory headroom simultaneously to guarantee holding all records.Because each mapping is all one section of continuous address, it is obviously inapplicable adopting the memory allocation method of operating system, while therefore creating index tree on mapping internal memory, needs a set of efficient memory management mechanism.
Common index machine is shaped with hash index, b-tree indexed etc., and wherein hash index efficiency is high, but because its monopolizing characteristic is not supported high-level efficiency range query, can not sort simultaneously, is not therefore used widely.And b-tree indexed is because its stability is high, support the features such as range query and sequence and be used widely.Tradition indexed mode often takes repeatedly the mode of access file to inquire about, and b-tree indexed utilizes the feature that the index structure number of plies is little, makes the number of times of access file relatively less.
summary of the invention
In order further to reduce number of times and the access time of access disk file, the present invention has proposed a kind of database main memory indexing means on the basis of B+ tree.The object of the invention is to reduce index file in data base querying and read in the time of internal memory, and further reduce the time of indexed search and the storage space of index structure, to realize more efficient database main memory index technology.
The present invention adopts following technical scheme for solving its technical matters:
A kind of database main memory indexing means, comprises the steps:
(1) when the major key that log-on data storehouse system is certain tables of data creates index, first adopt memory-mapped mode that newly-built index file is mapped in internal memory, on the first address of mapping region of memory, create indexing head structure;
(2) create after indexing head structure, read in index key value record according to tables of data, create index structure in this mapping area, by index data synchronous refresh in index file; When the index structure of region of memory is carried out to attended operation, need dynamically update indexing head information;
(3) while exiting Database Systems, remove memory-mapped, can realize on peripheral hardware disk and preserve index structure; Again index file is mapped into internal memory when restarting when system is inquired about, directly complete index structure is imported in internal memory, the high-level efficiency that realizes index structure imports and derives.
The index structure of region of memory is being carried out to attended operation, comprising:
When distribution and releasing memory, adopt chain type memory management mechanism, all free blocks are stringed together by the form of chained list, in the time being index record application memory headroom, need to utilize the pointer that points to free block in indexing head, the first node of its sensing free block chained list, the tail node of this free time chained list is positioned at position corresponding to side-play amount; When storage allocation, until find and can hold the big or small free block of application, upgrade free block chained list according to next traversal free block chained list in the time of Free up Memory, this piece of initialization also travels through chained list, find corresponding position and insert, observe whether adjacent with other pieces, if adjacent merging.
In the time of application memory headroom, also comprise:
Along with the insertion of index record, there is the situation of default file size deficiency, one threshold value is set in indexing head, when the ratio that accounts for file when file unassigned zone is less than this threshold value, disconnect mapping, according to dilatation number of times to index file dilatation, in the time that dilatation number of times is larger, each dilatation ratio is just larger, when certain number of times is counted in dilatation time, is made as definite value; After to be expanded completing, the index file after dilatation is remapped into internal memory.
A kind of database main memory indexing means, comprises the following steps:
Except supporting unique index, create the situation of index for the higher key assignments of repetition rate, in conjunction with the feature of hash and B tree, adopt a kind of improved b-tree indexed structure to support repeat key, this index structure adopts row indications corresponding to storage of linked list repeat key;
While carrying out attended operation, first will find the leaf node that interdependent node is corresponding, first check whether leaf node exists this key assignments when insertion, if do not exist, application internal memory is inserted into this node in this index tree; The row indications of node is inserted in the chained list that this key-value pair answers if exist, when deletion, the rower of first deleting in the corresponding chained list of key assignments shows token record, and whether be empty, this key assignments is deleted from index tree if it is empty to also Free up Memory if reexamining this chained list; Inquiry and modification are all in leaf node, to find corresponding key, in corresponding chained list, search and amendment record.
Beneficial effect of the present invention is as follows:
Obviously reduce the time that index file reads in internal memory and writes back, reduce the number of times of access index file, realizing efficient index structure imports and derives, effectively utilize storage space, support repeat key, in the time that constantly increasing, index record can, dynamically to index file dilatation, realize the inquiry of high-level efficiency and stable performance simultaneously.
accompanying drawing explanation
Fig. 1 is principle schematic of the present invention.
Fig. 2 is the schematic flow sheet of index file mapping.
Fig. 3 is the B tree construction schematic diagram with indexing head.
Fig. 4 improves b-tree indexed leaf node schematic diagram.
Fig. 5 is index file dynamic growth schematic diagram.
embodiment
Below in conjunction with accompanying drawing, the invention is described in further details.
Adopt traditional file read-write mode to carry out frequent operation for large file and can waste a large amount of time, the present invention adopts memory-mapped mode to store index file and solves this problem, and has created the data structure that one is called indexing head (index head) and manage mapping internal memory and index structure.
B-tree indexed support is carried out index to repeat key, in the time that the repetition rate of key is higher, tree is inserted into repeat key in index tree traditional B, efficiency is very low, the present invention has proposed a kind of improvement B tree of supporting repeat key in conjunction with hash algorithm, and the complexity that can effectively reduce storage space and tree guarantees query performance simultaneously.
When in the maintenance in later stage, along with index record progressively inserts, index file has and exceedes default big or small danger, the present invention adopts a kind of index file dynamic growth mechanism to carry out dilatation, to guarantee that having sufficient space holds index record, there will not be spillover, do not waste storage space, guaranteed like this expansibility that system is good simultaneously.
Memory-mapped (memory map) technology can be shone upon part or all of a file in internal memory.As shown in Figure 1, index file is all mapped in internal memory, and mapping scope is file size, can directly carry out with pointer like this to the operation of file, and operating system can flush to internal storage data in corresponding file, to realize dynamically updating of index file.
The technology that embodiment of the present invention adopts is exactly memory-mapped technology, adopt mmap () function that index file is mapped into internal memory according to certain way and size, obtain the base address of mapping, be mapped into after internal memory without visiting again index file, only need the index internal memory of access map, can save like this index file access time.After all operations completes and logs off, use munmap () function by this section of internal memory index file that remaps back, complete index upgrade.The address that is mapped to internal memory due to index file can be set to fixing and two kinds of patterns of system assignment, once consider fixing start address, in the time that this section of internal memory taken by other processes of system in advance, the portability of system can not be guaranteed, so the present invention adopts the mode of system assignment, the address of so each mapping is not identical, so the pointer in node can not be stored true address, but side-play amount corresponding to memory address, this side-play amount is the poor of memory address and base address (base_ptr); When access pointer indication variable, need to add base address.
Below by Figure of description, to each embodiment, the present invention will be described.
1) embodiment mono-
Embodiments of the invention one have been introduced a kind of method of index file importing internal memory, and concrete steps flow process as shown in Figure 2, comprising:
Whether A, inspection there is index, perform step B if exist; Otherwise execution step G;
B, record number according to tables of data and estimate index file size, and according to the index file of table name, the corresponding size of row name creation;
C, according to certain read-write mode, set document misregistration amount and mapping length index file is mapped into internal memory, mapping mode is taked Random Maps mode, by system Random assignment internal memory, location comes back to the base after having shone upon;
D, on base address, create indexing head, carry out initialization, by internal memory free block corresponding pointed side-play amount;
E, obtain successively index record according to data recording;
If F Article 1 record, traversal chained list application space creates index tree; Otherwise traversal chained list application space is inserted into node in index tree, specifically refers to the inserting step in embodiment bis-;
G, inquire about, delete, revise, the operation such as insertion;
H, log off, remove memory-mapped.
2) embodiment bis-
Embodiments of the invention two as shown in Figure 3,4, improved index structure and B data tree structure are shown, Fig. 3 represents improved b-tree indexed structure, it has comprised indexing head, wherein indexing head has pointed to respectively root node, leaf chain heading node and tail node, so just can find fast root node, inserts, deletes, inquires about, the operation such as modification, can obtain minimum and maximum key assignments index record, carry out aggregation operator to facilitate simultaneously.The concrete improvement part of B tree of the present invention as shown in Figure 4, it mainly improves the storage of duplicate key value, the key assignments record of repetition is stored in the row indications chained list that key-value pair answers, so just can avoid this record to be inserted in whole index tree, and the complex operations such as avoid dividing.
Embodiment bis-mainly comprises insertion, deletion, inquiry and revises four steps, as follows:
1, update step
The index record (key, rowid) that A, input are inserted into;
B, read root node root according to indexing head;
C, traversal index tree carry out key assignments key and relatively find corresponding leaf node;
If D finds leaf node, execution step E; Otherwise this tree is empty, creates index tree according to (key, rowid);
E, in this leaf node, find whether there is a key;
F, if there is no, execution step G; Otherwise execution step I;
G, be inserted in this node, check whether the node after inserting overflows, be inserted in father node if overflow, perform step H; Otherwise execution step J;
Whether H, inspection father node overflow, if overflow the operation of dividing and being inserted into father node, progressively iteration; Otherwise execution step J;
I, travel through this key assignments for chained list inquiry whether there is row indications rowid, if without it being inserted according to the order of sequence in the chained list that key assignments key points to, otherwise insert unsuccessfully;
J, return to root node, upgrade indexing head.
2, deletion action step
A, input index record to be deleted (key, rowid);
B, read root node root according to indexing head;
C, traverse tree carry out key assignments key and relatively find corresponding leaf node;
If D finds respective leaves node, execution step E; Otherwise represent that tree, for empty, exits;
E, the key that finds key value in this leaf node, if find, execution step F; Otherwise do not find respective nodes, delete unsuccessfully, exit;
F, in corresponding chained list, search row indications rowid, delete if find; Otherwise delete unsuccessfully, exit;
Whether the chained list that G, inspection key assignments key are corresponding is empty, if it is empty by key assignments key knot removal;
H, successfully delete index record, upgrade indexing head.
3, query manipulation step
A, input index key assignments key to be deleted;
B, read root node root according to indexing head;
C, traverse tree carry out key assignments key and relatively find corresponding leaf node;
If D finds leaf node, execution step E; Otherwise tree, for empty, exits;
E, the key that finds key value in this leaf node, if find to return the chained list first address of its sensing; Otherwise exit, do not find respective record.
4, retouching operation step
A, input index record to be modified (key, rowid_old, rowed_new);
B, read root node root according to indexing head;
C, traverse tree carry out key assignments key and relatively find corresponding leaf node;
If D finds leaf node, execution step E; Otherwise tree, for empty, exits;
E, the key that finds key value in this leaf node, if find execution step F, otherwise exit;
F, in chained list corresponding to key, search old row indications rowid_old, after finding, revise rowid_old and be new indications rowid_new, successfully modified, upgrade indexing head; Otherwise do not exist, revise unsuccessfully and exit.
3) embodiment tri-
Embodiment tri-provides a kind of index file dynamic capacity-expanding mechanism, and the steps flow chart of the method as shown in Figure 5, comprising:
Whether the ratio that A, the difference that checks unallocated memory size free_len and size to be allocated account for total size is less than the threshold value (5%) defining in indexing head;
If B is less than, execution step C; Otherwise execution step G;
C, disconnection mapping, write back disk file;
D, obtain the dilatation number of times extend_time that defines in indexing head and index file size index_file_len at present;
E, computation index file dilatation size;
Dilatation size extend_size is with dilatation number of times extend_time and two factors of index file size index_file_len are relevant at present:
If dilatation number of times is no more than 2 times
Dilatation size equals index file length * (2* dilatation number of times+1) * 10%;
Otherwise
Dilatation size equals index file length * 50%;
After F, dilatation, remap index file to internal memory;
G, proceed associative operation.

Claims (4)

1. a database main memory indexing means, is characterized in that comprising the steps:
(1) when the major key that log-on data storehouse system is certain tables of data creates index, first adopt memory-mapped mode that newly-built index file is mapped in internal memory, on the first address of mapping region of memory, create indexing head structure;
(2) create after indexing head structure, read in index key value record according to tables of data, create index structure in this mapping area, by index data synchronous refresh in index file; When the index structure of region of memory is carried out to attended operation, need dynamically update indexing head information;
(3) while exiting Database Systems, remove memory-mapped, can realize on peripheral hardware disk and preserve index structure; Again index file is mapped into internal memory when restarting when system is inquired about, directly complete index structure is imported in internal memory, the high-level efficiency that realizes index structure imports and derives.
2. a kind of database main memory indexing means as claimed in claim 1, is characterized in that the index structure of region of memory carrying out attended operation, comprising:
When distribution and releasing memory, adopt chain type memory management mechanism, all free blocks are stringed together by the form of chained list, in the time being index record application memory headroom, need to utilize the pointer that points to free block in indexing head, the first node of its sensing free block chained list, the tail node of this free time chained list is positioned at position corresponding to side-play amount; When storage allocation, until find and can hold the big or small free block of application, upgrade free block chained list according to next traversal free block chained list in the time of Free up Memory, this piece of initialization also travels through chained list, find corresponding position and insert, observe whether adjacent with other pieces, if adjacent merging.
3. a kind of database main memory indexing means as claimed in claim 2, is characterized in that, in the time of application memory headroom, also comprises:
Along with the insertion of index record, there is the situation of default file size deficiency, one threshold value is set in indexing head, when the ratio that accounts for file when file unassigned zone is less than this threshold value, disconnect mapping, according to dilatation number of times to index file dilatation, in the time that dilatation number of times is larger, each dilatation ratio is just larger, when certain number of times is counted in dilatation time, is made as definite value; After to be expanded completing, the index file after dilatation is remapped into internal memory.
4. a database main memory indexing means, is characterized in that, comprises the following steps:
Except supporting unique index, create the situation of index for the higher key assignments of repetition rate, in conjunction with the feature of hash and B tree, adopt a kind of improved b-tree indexed structure to support repeat key, this index structure adopts row indications corresponding to storage of linked list repeat key;
While carrying out attended operation, first will find the leaf node that interdependent node is corresponding, first check whether leaf node exists this key assignments when insertion, if do not exist, application internal memory is inserted into this node in this index tree; The row indications of node is inserted in the chained list that this key-value pair answers if exist, when deletion, the rower of first deleting in the corresponding chained list of key assignments shows token record, and whether be empty, this key assignments is deleted from index tree if it is empty to also Free up Memory if reexamining this chained list; Inquiry and modification are all in leaf node, to find corresponding key, in corresponding chained list, search and amendment record.
CN201410063636.0A 2014-02-25 2014-02-25 Database primary memory indexing method Pending CN103823865A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410063636.0A CN103823865A (en) 2014-02-25 2014-02-25 Database primary memory indexing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410063636.0A CN103823865A (en) 2014-02-25 2014-02-25 Database primary memory indexing method

Publications (1)

Publication Number Publication Date
CN103823865A true CN103823865A (en) 2014-05-28

Family

ID=50758929

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410063636.0A Pending CN103823865A (en) 2014-02-25 2014-02-25 Database primary memory indexing method

Country Status (1)

Country Link
CN (1) CN103823865A (en)

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104268146A (en) * 2014-08-21 2015-01-07 南京邮电大学 Static B+-tree index method suitable for analytic applications
CN104881441A (en) * 2015-05-12 2015-09-02 福建天晴数码有限公司 File receiving and storing method based on local area network, checking method and system
CN104899297A (en) * 2015-06-08 2015-09-09 南京航空航天大学 Hybrid index structure with storage perception
CN105843750A (en) * 2016-03-22 2016-08-10 深圳森格瑞通信有限公司 Linear data structure query efficiency improving method
CN106326397A (en) * 2016-08-19 2017-01-11 东软集团股份有限公司 Method and device for generating index file
CN106547805A (en) * 2015-09-23 2017-03-29 北京奇虎科技有限公司 The method and apparatus of optimization database index
CN106776702A (en) * 2016-11-11 2017-05-31 北京奇虎科技有限公司 A kind of method and apparatus for processing the index in master-slave mode Database Systems
CN106776811A (en) * 2016-11-23 2017-05-31 李天� data index method and device
CN106980578A (en) * 2017-04-01 2017-07-25 广东浪潮大数据研究有限公司 A kind of memory block management method and system
CN106991102A (en) * 2016-01-21 2017-07-28 腾讯科技(深圳)有限公司 The processing method and processing system of key-value pair in inverted index
CN107315745A (en) * 2016-04-26 2017-11-03 北京京东尚科信息技术有限公司 A kind of personal letter storage method and system
CN107885489A (en) * 2017-10-25 2018-04-06 中国信息通信研究院 A kind of method and system of quick detection real name registration data index
CN108052643A (en) * 2017-12-22 2018-05-18 北京奇虎科技有限公司 Date storage method, device and storage engines based on LSM Tree structures
CN108984780A (en) * 2018-07-25 2018-12-11 郑州云海信息技术有限公司 Based on the method and apparatus for supporting duplicate key value data tree structure management data
CN109583861A (en) * 2018-11-23 2019-04-05 中科驭数(北京)科技有限公司 Data compression method, access method and system in key-value database
CN110781101A (en) * 2019-10-25 2020-02-11 苏州浪潮智能科技有限公司 One-to-many mapping relation storage method and device, electronic equipment and medium
WO2020041950A1 (en) * 2018-08-27 2020-03-05 袁振南 Data update method, device, and storage device employing b+ tree indexing
CN111226205A (en) * 2017-08-31 2020-06-02 美光科技公司 KVS tree database
CN113434489A (en) * 2021-08-26 2021-09-24 西安热工研究院有限公司 Real-time database online capacity expansion method, system, equipment and storage medium

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040193632A1 (en) * 2003-03-27 2004-09-30 Mccool Michael Computer implemented compact 0-complete tree dynamic storage structure and method of processing stored data
CN101639848A (en) * 2009-06-01 2010-02-03 北京四维图新科技股份有限公司 Spatial data engine and method applying management spatial data thereof
CN101783740A (en) * 2009-01-21 2010-07-21 大唐移动通信设备有限公司 Method and device for managing message file
CN102231121A (en) * 2011-07-25 2011-11-02 北方工业大学 Method for rapidly extracting massive data files in parallel based on memory mapping
CN102385623A (en) * 2011-10-25 2012-03-21 曙光信息产业(北京)有限公司 Catalogue access method in DFS (distributed file system)

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040193632A1 (en) * 2003-03-27 2004-09-30 Mccool Michael Computer implemented compact 0-complete tree dynamic storage structure and method of processing stored data
CN101783740A (en) * 2009-01-21 2010-07-21 大唐移动通信设备有限公司 Method and device for managing message file
CN101639848A (en) * 2009-06-01 2010-02-03 北京四维图新科技股份有限公司 Spatial data engine and method applying management spatial data thereof
CN102231121A (en) * 2011-07-25 2011-11-02 北方工业大学 Method for rapidly extracting massive data files in parallel based on memory mapping
CN102385623A (en) * 2011-10-25 2012-03-21 曙光信息产业(北京)有限公司 Catalogue access method in DFS (distributed file system)

Cited By (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104268146A (en) * 2014-08-21 2015-01-07 南京邮电大学 Static B+-tree index method suitable for analytic applications
CN104881441B (en) * 2015-05-12 2018-11-06 福建天晴数码有限公司 File reception and storage method, inspection method based on LAN and system
CN104881441A (en) * 2015-05-12 2015-09-02 福建天晴数码有限公司 File receiving and storing method based on local area network, checking method and system
CN109376156A (en) * 2015-06-08 2019-02-22 南京航空航天大学 Read the method with the hybrid index of storage perception
CN109376156B (en) * 2015-06-08 2021-09-21 南京航空航天大学 Method for reading hybrid index with storage awareness
CN104899297B (en) * 2015-06-08 2019-02-26 南京航空航天大学 Create the method with the hybrid index of storage perception
CN104899297A (en) * 2015-06-08 2015-09-09 南京航空航天大学 Hybrid index structure with storage perception
CN106547805A (en) * 2015-09-23 2017-03-29 北京奇虎科技有限公司 The method and apparatus of optimization database index
CN106991102A (en) * 2016-01-21 2017-07-28 腾讯科技(深圳)有限公司 The processing method and processing system of key-value pair in inverted index
CN106991102B (en) * 2016-01-21 2021-06-08 腾讯科技(深圳)有限公司 Processing method and processing system for key value pairs in inverted index
CN105843750A (en) * 2016-03-22 2016-08-10 深圳森格瑞通信有限公司 Linear data structure query efficiency improving method
CN107315745A (en) * 2016-04-26 2017-11-03 北京京东尚科信息技术有限公司 A kind of personal letter storage method and system
CN106326397A (en) * 2016-08-19 2017-01-11 东软集团股份有限公司 Method and device for generating index file
CN106776702A (en) * 2016-11-11 2017-05-31 北京奇虎科技有限公司 A kind of method and apparatus for processing the index in master-slave mode Database Systems
CN106776702B (en) * 2016-11-11 2021-03-05 北京奇虎科技有限公司 Method and device for processing indexes in master-slave database system
CN106776811A (en) * 2016-11-23 2017-05-31 李天� data index method and device
CN106980578A (en) * 2017-04-01 2017-07-25 广东浪潮大数据研究有限公司 A kind of memory block management method and system
CN111226205A (en) * 2017-08-31 2020-06-02 美光科技公司 KVS tree database
CN111226205B (en) * 2017-08-31 2021-08-31 美光科技公司 KVS tree database
CN107885489A (en) * 2017-10-25 2018-04-06 中国信息通信研究院 A kind of method and system of quick detection real name registration data index
CN107885489B (en) * 2017-10-25 2021-05-04 中国信息通信研究院 Method and system for rapidly detecting real-name registration data indexes
CN108052643A (en) * 2017-12-22 2018-05-18 北京奇虎科技有限公司 Date storage method, device and storage engines based on LSM Tree structures
CN108052643B (en) * 2017-12-22 2021-02-23 北京奇虎科技有限公司 Data storage method and device based on LSM Tree structure and storage engine
CN108984780A (en) * 2018-07-25 2018-12-11 郑州云海信息技术有限公司 Based on the method and apparatus for supporting duplicate key value data tree structure management data
CN108984780B (en) * 2018-07-25 2021-10-22 郑州云海信息技术有限公司 Method and device for managing disk data based on data structure supporting repeated key value tree
WO2020041950A1 (en) * 2018-08-27 2020-03-05 袁振南 Data update method, device, and storage device employing b+ tree indexing
CN109583861A (en) * 2018-11-23 2019-04-05 中科驭数(北京)科技有限公司 Data compression method, access method and system in key-value database
CN110781101A (en) * 2019-10-25 2020-02-11 苏州浪潮智能科技有限公司 One-to-many mapping relation storage method and device, electronic equipment and medium
CN113434489A (en) * 2021-08-26 2021-09-24 西安热工研究院有限公司 Real-time database online capacity expansion method, system, equipment and storage medium

Similar Documents

Publication Publication Date Title
CN103823865A (en) Database primary memory indexing method
US8799284B2 (en) Method for automated scaling of a massive parallel processing (MPP) database
CN102129458B (en) Method and device for storing relational database
CN104484471B (en) A kind of implementation method of high-performance data storage engines
CN102521334B (en) Data storage and query method based on classification characteristics and balanced binary tree
CN105117415B (en) A kind of SSD data-updating methods of optimization
CN103838853B (en) Mixed file system based on different storage media
CN105117417B (en) A kind of memory database Trie tree indexing means for reading optimization
CN103914483B (en) File memory method, device and file reading, device
CN105320775A (en) Data access method and apparatus
CN104133661A (en) Multi-core parallel hash partitioning optimizing method based on column storage
CN103473298B (en) Data archiving method and device and storage system
CN103020255A (en) Hierarchical storage method and hierarchical storage device
CN105159616A (en) Disk space management method and device
CN105975587A (en) Method for organizing and accessing memory database index with high performance
CN110188108A (en) Date storage method, device, system, computer equipment and storage medium
CN107783988A (en) The locking method and equipment of a kind of directory tree
US6745198B1 (en) Parallel spatial join index
US8682872B2 (en) Index page split avoidance with mass insert processing
WO2016188280A1 (en) Write-in method and device for database sub-tables
CN107273443B (en) Mixed indexing method based on metadata of big data model
US20210056087A1 (en) Method and apparatus for providing efficient indexing and computer program included in computer readable medium therefor
CN103229164B (en) Data access method and device
JP6006740B2 (en) Index management device
CN110597805A (en) Efficient novel memory index structure processing method

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20140528