CN102654863A - Real-time database history data organizational management method - Google Patents
Real-time database history data organizational management method Download PDFInfo
- Publication number
- CN102654863A CN102654863A CN2011100497608A CN201110049760A CN102654863A CN 102654863 A CN102654863 A CN 102654863A CN 2011100497608 A CN2011100497608 A CN 2011100497608A CN 201110049760 A CN201110049760 A CN 201110049760A CN 102654863 A CN102654863 A CN 102654863A
- Authority
- CN
- China
- Prior art keywords
- page
- leaf
- data
- state
- node
- 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
Links
Images
Abstract
The invention discloses a real-time database history data organizational management method. The method mainly comprises: real-time database history buffer area management and real-time database history file indexing modes. Through the history data organizational management method disclosed by the invention, a memory space with a designated size is allocated to the history data buffer area in advance, and the buffer area reading/writing management is performed in the memory space in a unified manner; the history data can be quickly written through history file management and the improved mode of indexing the history data by use of b+ tree in combination with proper buffer policy, page request mode and dirty page cleaning mode; and a history data index is established at relatively low cost at the time of writing so that the search for history data is very efficient. Meanwhile, by adopting a history data buffer area with a fixed size and a management mode of unifying reading/writing and buffering, the history data management model is greatly simplified, and the stability of the whole real-time database history data management module is improved.
Description
Technical field
The present invention relates to the real-time data base technology, be meant a kind of method that the real-time data base historical data is managed especially.
Background technology
Real-time data base is the product that database technology combines with real-time system.Real-time data base has a wide range of applications in workflow industry (petrochemical industry, electric power, iron and steel).It can be used for automatic collection, storage and the supervision of production run, and data for many years that also can online each technological process point of storage are cores of the MES of enterprise.But real-time data base is not isolated the existence, and it has constituted SIS jointly with other software and hardware spares, the system that ERP etc. are complicated, and this wherein relates to DCS, PLC, fieldbus, relational database, a series of complex conception such as human-machine interface configuration.Can therefore, the quality of real-time data base not only depends on himself, also depend on real-time data base for industry energy conservation reduces discharging, create maximum economic benefits and social benefit well with above-mentioned software and hardware collaborative work.
The main characteristic of real-time data base is exactly that its data and task all have explicit timing restriction; But because real-time data base is mainly used in workflow industry; Scale is bigger and workflow industry is counted usually, in the real-time that requires, also will accomplish the storage for a large amount of historical datas, so that ex-post analysis and emergency review; Therefore how efficiently the historical data of real-time data base to be carried out organization and administration, just become the key factor of decision real-time data base performance quality.
In the management of historical data tissue, there are problems in present real-time data base product; Being mainly reflected in (1) adopts the small documents of a plurality of fixed sizes to carry out the storage of historical data; Make the management process complicacy very of history file, the data directory mode of performance also not high (2) history data file is comparatively simple, lacks dirigibility; Historical data look into the neglect history buffer management of file of institute's aspect performance not good (3); Too rely on the file buffering mechanism use operating system self, can not effectively adapt to the rule of real-time data base historical data visit, thereby efficient is not high aspect the writing and read of historical data.
Summary of the invention
In view of this; Fundamental purpose of the present invention is to provide a kind of method of real-time data base historical data management, when promoting the historical data retrieval rate, has guaranteed that the high speed of a large amount of historical datas writes; This way to manage also has outstanding performance aspect system stable.
In order to achieve the above object, in the design of this method, introduced following mechanism:
Utilize the some catalogue to carry out all measuring point management of information.
Return the buffer zone page address of containing the specified file page data through syllogic page request mechanism.
Utilize state-chain-table, the page in the buffer zone is classified, the aspect skips leaf and handles and dirty page or leaf cleaning.
Improved b+ tree data are inserted and division mechanism.
The regular dirty page or leaf Prune Policies of base state chained list.
Preferably, the concrete mode of said some directory information management is:
The index of each measuring point all is based on a kind of improved b+ tree, need in measuring point catalogue buffer zone, write down the data page the earliest of the corresponding b+ tree of measuring point correspondence, data page the latest, the root page.
Through data page the earliest, the user can find the data the earliest of a certain measuring point fast, then through the chained list of data page, travels through all historical datas of measuring point fast.And pass through data page the latest, when inserting new historical data, system can locate the insertion position of new data fast, and saves the cost of searching the insertion position.Because generally speaking, user's the inquiry of historical data request randomness is very strong, therefore, must in a catalogue, write down each and improve b+ tree root node so that locate fast the page of data that will search.
Preferably, said syllogic page request mechanism specifically comprises:
Phase one is searching the stage of the history buffer page, through given document number and page number, through the calculating of hash function; System can judge fast whether the page of being asked is present in the middle of the history buffer; If exist, then return the address of the buffer zone page, if do not exist; Then get into second stage of page request mechanism, the processing stage of skipping leaf.
The processing stage skipping leaf.System need find a new buffer zone page fast for the file page that skips leaf; So that be written into the data page in the middle of the disk; This new buffer zone page should be idle not to be written into any data in magnetic disk or in short-term, the file data of its representative can not had access to once more.Realize this purpose, just need be by state-chain-table,
Three phases is after the page access of request is accomplished, the subsequent treatment stage of the page.This stage need divide two kinds of situation to handle, and a kind of is page read request subsequent treatment, and a kind of is that the page is write the request subsequent treatment.
Preferably, the said hash function mechanism of searching further is:
With given document number and page number, form a long, hash function, can be mapped to the long of representing specified page on a certain address through hash function for getting complementary function, and then through chain address method, the processing that conflicts.Can carry out searching of buffer zone page info fast through this mode.
If search page info less than appointment, explain and taken place to skip leaf that need change the processing of skipping leaf over to, the process result that skips leaf is that the structure with a new representing pages information is inserted on the corresponding hash address, the process of insertion with search similar.
Preferably, the said treatment mechanism that skips leaf further is:
In case in hash table, can not find the page information of appointment, just need scan one by one state-chain-table, therefrom select a suitable buffer zone page as the object that is written into data in magnetic disk again then.
The sequencing of scanning is followed successively by:
IDLE-＞C_IL-＞C_II-＞D_II-＞D_IL-＞D_AL
Preferably, said page subsequent treatment mechanism further is:
According to the access type of the page, whether the visit overshoot skips leaf, and the dbjective state chained list that provides of visitor, confirms the end-state after some page access.
The access type of the page mainly is divided into to be read the page and writes the page, for reading the page, if in the process of page request, does not skip leaf; Can directly return, otherwise, the dbjective state chain table number that need provide according to the visitor; Carry out the migration of page status, for writing the page, no matter in the access process of the page; Whether taken place to skip leaf, all need according to the dbjective state chain table number that the visitor provides, the node of representing pages information has been moved into the state-chain-table of correspondence.Whether the dbjective state chained list that the visitor provides is semantic according to the upper strata in the page access process, consistent such as the data of the page, and perhaps whether node has been write and completely come to confirm.In the middle of the process of page status migration, the present residing state-chain-table of the page how, the dbjective state chained list that provides with the visitor is consistent, then need not carry out the migration of page status.
Preferably, said state-chain-table mechanism further is:
According to the kind of the page, data consistency, and liveness are reasonably divided the page in the buffer zone, thereby conveniently skip leaf processing, dirty page or leaf clean-up task accurately judge the page state, utilize each buffer zone page efficiently.The state-chain-table that exists in the system has:
(1) idle chained list (abbreviating STATE_IDLE as)
(2) inactive leaf node chained list (totally) (abbreviating STATE_C_IL as)
(3) inactive leaf node chained list (dirty) (abbreviating STATE_D_IL as)
(4) enliven leaf node chained list (dirty) (abbreviating STATE_D_AL as)
(5) enliven internal node chained list (dirty) (abbreviating STATE_D_AI as)
(6) inactive internal node chained list (totally) (abbreviating STATE_C_II as)
(7) inactive internal node chained list (dirty) (abbreviating STATE_D_II as)
Preferably, said data insertion and splitter are made as:
Whenever nodes records or key word write full after, the division of node all is chosen in last key word place and carries out.For a leaf node with n bar record, two nodes after the division, one has n-1 key word; One has last 1 key word, simultaneously last key word is duplicated portion, is inserted in the middle of the father node; For a internal node with n bar record; Two nodes after the division, one has n-2 key word, and one has last 1 key word.The second-to-last key word then inserts in the middle of the father node.
Preferably, said regular dirty page or leaf Prune Policies is:
Whenever at a distance from section regular time; The buffer zone page data of not visited on D_II and two state-chain-tables of D_IL is write back in the middle of the disk; Then with among the corresponding buffer region page immigration state-chain-table IDLE; So that skip leaf in the process of next page request, the processing procedure that skips leaf can find the reusable buffer zone page fast.
Description of drawings
Shown in Figure 1 is general frame figure of the present invention;
Fig. 2 is the synoptic diagram of syllogic page request mechanism among the embodiment one.
Fig. 3 is the leaf node divisional mode of the b+ tree after improving in the real-time example two.
Fig. 4 is the internal node divisional mode of the b+ tree after improving in the real-time example two.
Embodiment
For making the object of the invention, technical scheme and advantage clearer, lift specific embodiment below, the present invention is done further detailed explanation.
As can beappreciated from fig. 1 come; The access request of any historical data all needs to find the corresponding page number of b+ tree node through a catalogue earlier; Then with page number as input, the page request mechanism through system finds the corresponding buffer region page; The read-write utilization of historical data be identical buffer zone; Therefore the page access in this buffer zone just can be divided into read access and write access, and what the index of history data file adopted in the system is the b+ tree after a kind of the improvement, different its divisional modes that are of setting with traditional b+ of b+ tree in this.It is that dirty page or leaf is handled that historical data when historical data is inserted in the buffer memory exchanges to important channel in the external memory, and the main path that deposit data is written into internal memory outside when the inquiry of historical data is the processing of skipping leaf.
Embodiment one:
Fig. 2 is the synoptic diagram of syllogic page request mechanism among the embodiment one.As shown in Figure 2, the process of page request mechanism mainly may further comprise the steps:
Step 101 is through hash function positioning webpage address.
After the user provides the document number and page number of request, it is formed a long.Through hash function, find corresponding hash address, along chained list scan page list item one by one; Whether the pgno field of contrast page table entry and given page number mate, and coupling is then returned the corresponding page address; After if scanning is accomplished, do not find corresponding page table entry, then turn to the processing of skipping leaf.
, should be noted that when inquiring a certain page table entry during enterprising line scanning in the hash address of correspondence; If page table entry matees, and the page of corresponding buffer zone in have other threads when visiting, should at first discharge locking to ltsh chain table; After postponing a period of time simultaneously; This page is conducted interviews, under the busy situation of systematic comparison, this way can reduce the possibility that conflict takes place again.
Step 102, scanning mode chained list IDLE → C_IL → C_II → D_II → D_IL → D_AL one by one.
Because the page table entry in the IDLE chained list is not represented any one file page, can not exert an influence for the utilization of these page table entries to other database tasks, therefore, the page table entry on the IDLE chained list has just become to skip leaf when handling, the first-selection when page table entry is searched.Processing procedure to the IDLE page table entry is also fairly simple, and the page number field that only need fill in page info gets final product in the ltsh chain table of insertion correspondence then.If the page table entry in the IDLE chained list exhausts, the processing procedure that skips leaf can be at C_IL, continues to search in these two state-chain-tables of CII; Owing to deposit in these two chained lists, be in the query script of historical data, clean () page table entry of generation; Also fairly simple to reusing of these page table entries, only need they be taken above original ltsh chain table, be inserted in the middle of the hash address of skipping leaf; Because these operations are all carried out, therefore can guarantee the high efficiency that skips leaf and handle in internal memory.
If still can not find the suitable page, the processing of skipping leaf can be transferred inquiry D_II, D_IL, these three chained lists of D_AL.
Step 103 finds the suitable buffer zone page, is written into file page.
And C_IL, the C_II chained list is different, D_II; D_IL, the page data on the D_AL chained list are " dirty ", that is to say; Before utilizing the buffer memory page again; Must corresponding data be write back earlier in the middle of the disk block, and then the content of the disk file page or leaf of being asked is loaded in the middle of the buffer zone page, with the data consistency of external memory in guaranteeing.When being written into the disk file page or leaf, what the requestor of the page was clear and definite knows that the own page of being asked is a blank page, so, can forbid being written into of data in magnetic disk through a flag; Can remove the process of a disk access like this from, by contrast, D_II; D_IL, the liveness of D_AL is the highest in three chained lists of D_AL, has only when other state-chain-tables perhaps do not have available page table entry for sky; Just can reuse the page on this chained list, go to this step if skip leaf to handle, the performance of system will produce tangible decline; Simultaneously, indicating that also buffer resource is about to exhaust, and should consider to increase the buffer zone of system.
Step 104 is judged the page access type and is skipped leaf situation, obtains the purpose state-chain-table.
Along with the division of the page in the historical data additive proccess (seeing the 5-3 joint), and the processing of skipping leaf in the page request process, the page of buffer zone constantly from a state-chain-table, is transferred to another state-chain-table.
Generally speaking, the requestor of page data can be according to the state of page data after the visit that is through with to the page, the clear and definite residing state of this page that provides, and just corresponding page table entry should get among that state-chain-table.For example; Be positioned at the page on the D_AL chained list; When the record number of the page did not reach the division requirement, after historical data was each time inserted, corresponding page table entry all should continue to reside in above the D_AL; In case the page data page, corresponding page table entry just should be transferred in the middle of the D_IL by the D_AL chained list.Again such as, when in the access process of the page, skipping leaf, which state-chain-table newly assigned page table entry should get into, the requestor of the page can provide naturally.Whether on the other hand, the upper strata semanteme is clearly for the read-write type of the page certainly, as for skipping leaf, can in the process that hash function is searched, know.
After having grasped these information, for reading the page,, do not skip leaf if in the process of page request; Can directly return, otherwise, the dbjective state chain table number that need provide according to the visitor; Carry out the migration of page status, for writing the page, no matter in the access process of the page; Whether taken place to skip leaf, all need according to the dbjective state chain table number that the visitor provides, the node of representing pages information has been moved into the state-chain-table of correspondence.
Step 105 moves into the page state-chain-table of appointment.
Because the state transition of the page relates to locking and unlocking operation of chained list, therefore time cost, when specified page state transition strategy, should reduce the number of times of page status migration than higher as far as possible under the prerequisite of guaranteed efficiency and correctness.
The state field of page table entry always with page table entry before residing state-chain-table consistent, whenever page access finishes, if skip leaf; Then read operation can not exert an influence to the state of the page, if taken place to skip leaf, will see whether given page status is consistent with the state field of page table entry; If it is consistent; Naturally need not carry out any operation,, just should carry out the state transition of chained list if inconsistent.
Embodiment two:
Embodiment two has described in the b+ tree after improving, the divisional mode of node.
Fig. 3 is the leaf node divisional mode of the b+ tree after improving in the real-time example two.
As shown in Figure 3, append under the pattern, the timestamp of new data is always maximum; Can write direct and deposit in the middle of the page of historical record the latest, and not need comparison, this page is positioned at the end of leaf node chained list; And be constantly to change, should write down the page number the latest of each measuring point in the database, after the page is write completely the latest; Just need carry out the division of leaf node, the division of leaf node is different with the division of internal node.
When carrying out data supplementing; After leaf node is write completely the latest; A record the latest wherein, the key word of that just writes an at last record can write in the middle of the father node, simultaneously; This record also can be copied in the middle of the newly assigned leaf node, and this newly assigned leaf node has also replaced original leaf node the latest with regard to having become new leaf node the latest.Since when carrying out the appending of historical data, the comparison that need not write down, and therefore, under the situation that leaf node does not divide, the expense of appending of wall scroll record is very little.
The b+ tree has only a leaf node initialized the time, in the time of the leaf node first division, need apply for new father node for it, and the level of b+ tree also increases thereupon simultaneously, and this is in the process of leaf node division, a more special place.
Fig. 4 is the internal node divisional mode of the b+ tree after improving in the real-time example two.
In the insertion process of historical data, internal node also is faced with in the ablation process of key word, the fragmentation problem of node.The key word of internal node, the key word that comes from fully in the above-mentioned leaf node fission process inserts.If singly see internal node, the key word of all internal nodes on b+ tree can not repeat, this be because, in the internal node division, have only moving of key word, do not have duplicating of key word.In addition, because b+ tree has level, that might once write down appends the node splitting that can cause on the many levels or even the division of root node.How to handle this iterative process well, be the internal node key word append with fission process in difficult point place.Fig. 4 has showed that the key word of an internal node appends and fission process.
This is once very typical node splitting process, and this wherein exists the division of the repeatedly division, particularly root node of node, with regard to the number of times of division; The node of lower floor, division frequent more, the node on upper strata, division slow more; The level of tree is high more, and node is not easy division more, in case root node divides; The height that means whole b+ tree adds 1, at this time, just needs to distribute new root node.
The above is merely preferred embodiment of the present invention, and is in order to restriction the present invention, not all within spirit of the present invention and principle, any modification of being done, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (9)
1. a real-time data base historical data is organized way to manage, it is characterized in that, this method comprises:
When carrying out the insertion of historical data, at first, find the page of data the latest of specifying measuring point through some catalogue buffer zone, after the page that navigates to appointment,, find the file page corresponding buffer region page through page request mechanism.Write historical data.
Writing in the middle of the process of historical data, if the full situation of data leaf has taken place, need carry out the division of page of data, if the full situation of index page has taken place corresponding father node, need proceed division, till root node.
The dirty page or leaf of buffer zone in the historical data ablation process can be write back disk by dirty page or leaf cleaning thread.
When carrying out the inquiry of historical data, still to find or be written into the buffer zone page of appointment through page request mechanism.No matter be reading or writing of the buffer zone page, the interruption of all might skipping leaf, and the migration of page status.
2. method according to claim 1 is characterized in that, the index of each measuring point all is based on a kind of improved b+ tree, need in measuring point catalogue buffer zone, write down the data page the earliest of the corresponding b+ tree of measuring point correspondence, data page the latest, the root page.
Through data page the earliest, the user can find the data the earliest of a certain measuring point fast, then through the chained list of data page, travels through all historical datas of measuring point fast.And pass through data page the latest, when inserting new historical data, system can locate the insertion position of new data fast, and saves the cost of searching the insertion position.Because generally speaking, user's the inquiry of historical data request randomness is very strong, therefore, must in a catalogue, write down each and improve b+ tree root node so that locate fast the page of data that will search.
3. method according to claim 1 is characterized in that, described page request mechanism comprises three main stages: the phase one is searching the stage of the history buffer page; Through given document number and page number, through the calculating of hash function, system can judge fast whether the page of being asked is present in the middle of the history buffer; If exist, then return the address of the buffer zone page, if do not exist; Then get into second stage of page request mechanism, the processing stage of skipping leaf.
The processing stage skipping leaf.System need find a new buffer zone page fast for the file page that skips leaf; So that be written into the data page in the middle of the disk; This new buffer zone page should be idle not to be written into any data in magnetic disk or in short-term, the file data of its representative can not had access to once more.Realize this purpose, just need be by state-chain-table,
Three phases is after the page access of request is accomplished, the subsequent treatment stage of the page.This stage need divide two kinds of situation to handle, and a kind of is page read request subsequent treatment, and a kind of is that the page is write the request subsequent treatment.
4. method according to claim 1 is characterized in that, described page request mode is specially:
With given document number and page number, form a long, hash function, can be mapped to the long of representing specified page on a certain address through hash function for getting complementary function, and then through chain address method, the processing that conflicts.Can carry out searching of buffer zone page info fast through this mode.
If search page info less than appointment, explain and taken place to skip leaf that need change the processing of skipping leaf over to, the process result that skips leaf is that the structure with a new representing pages information is inserted on the corresponding hash address, the process of insertion with search similar.
5. method according to claim 3; It is characterized in that the process of handling that skips leaf does, in case in hash table, can not find the page information of appointment; Just need scan one by one, therefrom select a suitable buffer zone page then as the object that is written into data in magnetic disk again to state-chain-table.The sequencing of scanning is followed successively by:
IDLE＞C_IL-＞C_II-＞D_II＞D_IL-＞D_AL
6. method according to claim 5 is characterized in that, the effect of state-chain-table is: according to the kind of the page; Data consistency; And liveness, the page in the buffer zone is reasonably divided, thereby conveniently skipped leaf processing; Dirty page or leaf clean-up task accurately judge the page state, utilize each buffer zone page efficiently.The state-chain-table that exists in the system has:
(1) idle chained list (abbreviating STATE_IDLE as)
(2) inactive leaf node chained list (totally) (abbreviating STATE_C_IL as)
(3) inactive leaf node chained list (dirty) (abbreviating STATE_D_IL as)
(4) enliven leaf node chained list (dirty) (abbreviating STATE_D_AL as)
(5) enliven internal node chained list (dirty) (abbreviating STATE_D_AI as)
(6) inactive internal node chained list (totally) (abbreviating STATE_C_II as)
(7) inactive internal node chained list (dirty) (abbreviating STATE_D_II as)
7. method according to claim 3 is characterized in that the effect of page request subsequent treatment is; Access type according to the page; Whether the visit overshoot skips leaf, and the dbjective state chained list that provides of visitor, confirms the end-state after some page access.
The access type of the page mainly is divided into to be read the page and writes the page, for reading the page, if in the process of page request, does not skip leaf; Can directly return, otherwise, the dbjective state chain table number that need provide according to the visitor; Carry out the migration of page status, for writing the page, no matter in the access process of the page; Whether taken place to skip leaf, all need according to the dbjective state chain table number that the visitor provides, the node of representing pages information has been moved into the state-chain-table of correspondence.Whether the dbjective state chained list that the visitor provides is semantic according to the upper strata in the page access process, consistent such as the data of the page, and perhaps whether node has been write and completely come to confirm.In the middle of the process of page status migration, the present residing state-chain-table of the page how, the dbjective state chained list that provides with the visitor is consistent, then need not carry out the migration of page status.
8. method according to claim 3 is characterized in that, the node splitting mode of b+ tree is different from traditional b+ tree; Traditional b+ tree is inserting under the recording key situation at random in order to guarantee, the integral body of tree has good balance all the time, in leaf node and root node division; Selection is cut apart from the middle part of node, and for native system b+ tree, because the key word of record is time corresponding and stabs; Therefore, the new key word that inserts increases progressively in order, in the balance that keeps tree; Make full use of disk space, the division of node all is chosen in last key word place and carries out.For a leaf node with n bar record, two nodes after the division, one has n-1 key word; One has last 1 key word, simultaneously last key word is duplicated portion, is inserted in the middle of the father node; For a internal node with n bar record; Two nodes after the division, one has n-2 key word, and one has last 1 key word.The second-to-last key word then inserts in the middle of the father node.
9. method according to claim 3; It is characterized in that the strategy of dirty page or leaf cleaning is, whenever at a distance from section regular time; The buffer zone page data of not visited on D_II and two state-chain-tables of D_IL is write back in the middle of the disk; With among the corresponding buffer region page immigration state-chain-table IDLE, so that skip leaf in the process of next page request, the processing procedure that skips leaf can find the reusable buffer zone page fast then.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100497608A CN102654863A (en) | 2011-03-02 | 2011-03-02 | Real-time database history data organizational management method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100497608A CN102654863A (en) | 2011-03-02 | 2011-03-02 | Real-time database history data organizational management method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102654863A true CN102654863A (en) | 2012-09-05 |
Family
ID=46730489
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011100497608A Pending CN102654863A (en) | 2011-03-02 | 2011-03-02 | Real-time database history data organizational management method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102654863A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020169A (en) * | 2012-11-27 | 2013-04-03 | 山东黄金矿业（莱州）有限公司 | Effectiveness and uniqueness processing method for electric data |
| CN103049396A (en) * | 2012-12-10 | 2013-04-17 | 浪潮(北京)电子信息产业有限公司 | Method and device for flushing data |
| CN107977417A (en) * | 2017-11-22 | 2018-05-01 | 江苏翼企云通信科技有限公司 | A kind of ultralight amount framework method accessed towards big data content high frequency |
| CN109669814A (en) * | 2018-12-10 | 2019-04-23 | 浪潮(北京)电子信息产业有限公司 | A kind of restoration methods of Metadata Service, device, equipment and readable storage medium storing program for executing |
| CN110456749A (en) * | 2019-08-13 | 2019-11-15 | 中国核动力研究设计院 | Based on nuclear power plant DCS platform engineering teacher's station software historical data processing method |
| CN111061802A (en) * | 2019-12-26 | 2020-04-24 | 宁波三星医疗电气股份有限公司 | Power data management processing method and device and storage medium |
| CN111198660A (en) * | 2019-12-26 | 2020-05-26 | 天津中科曙光存储科技有限公司 | B + tree traversal method and device |
| CN113194124A (en) * | 2021-04-19 | 2021-07-30 | 西安热工研究院有限公司 | Distributed real-time database-based DCS background multi-node inter-node directory synchronization method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1928881A (en) * | 2006-09-26 | 2007-03-14 | 南京擎天科技有限公司 | Computer data security protective method |
| CN101593220A (en) * | 2008-05-28 | 2009-12-02 | 北京华大九天软件有限公司 | A kind of management method of very large scale integrated circuit layout data |
| EP2282266A1 (en) * | 2009-07-24 | 2011-02-09 | Apple Inc. | Index cache tree |
-
2011
- 2011-03-02 CN CN2011100497608A patent/CN102654863A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1928881A (en) * | 2006-09-26 | 2007-03-14 | 南京擎天科技有限公司 | Computer data security protective method |
| CN101593220A (en) * | 2008-05-28 | 2009-12-02 | 北京华大九天软件有限公司 | A kind of management method of very large scale integrated circuit layout data |
| EP2282266A1 (en) * | 2009-07-24 | 2011-02-09 | Apple Inc. | Index cache tree |
Non-Patent Citations (1)
| Title |
|---|
| 杜冀秦: "基于B+树索引的实时数据库缓冲区管理算法", 《工程技术》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020169A (en) * | 2012-11-27 | 2013-04-03 | 山东黄金矿业（莱州）有限公司 | Effectiveness and uniqueness processing method for electric data |
| CN103049396A (en) * | 2012-12-10 | 2013-04-17 | 浪潮(北京)电子信息产业有限公司 | Method and device for flushing data |
| CN103049396B (en) * | 2012-12-10 | 2015-10-28 | 浪潮(北京)电子信息产业有限公司 | The method of writing with a brush dipped in Chinese ink of data and device |
| CN107977417A (en) * | 2017-11-22 | 2018-05-01 | 江苏翼企云通信科技有限公司 | A kind of ultralight amount framework method accessed towards big data content high frequency |
| CN109669814A (en) * | 2018-12-10 | 2019-04-23 | 浪潮(北京)电子信息产业有限公司 | A kind of restoration methods of Metadata Service, device, equipment and readable storage medium storing program for executing |
| CN110456749A (en) * | 2019-08-13 | 2019-11-15 | 中国核动力研究设计院 | Based on nuclear power plant DCS platform engineering teacher's station software historical data processing method |
| CN110456749B (en) * | 2019-08-13 | 2022-02-18 | 中核控制系统工程有限公司 | Nuclear power plant DCS platform based engineer station software historical data processing method |
| CN111061802A (en) * | 2019-12-26 | 2020-04-24 | 宁波三星医疗电气股份有限公司 | Power data management processing method and device and storage medium |
| CN111198660A (en) * | 2019-12-26 | 2020-05-26 | 天津中科曙光存储科技有限公司 | B + tree traversal method and device |
| CN113194124A (en) * | 2021-04-19 | 2021-07-30 | 西安热工研究院有限公司 | Distributed real-time database-based DCS background multi-node inter-node directory synchronization method |
| CN113194124B (en) * | 2021-04-19 | 2022-10-25 | 西安热工研究院有限公司 | Distributed real-time database-based DCS background multi-node directory synchronization method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102654863A (en) | Real-time database history data organizational management method | |
| CN100458779C (en) | Index and its extending and searching method | |
| CN105320775B (en) | The access method and device of data | |
| CN1026631C (en) | Method of concurrency access for useing indexed tree to data base record | |
| CN103412803B (en) | The method and device that data are recovered | |
| CN102033954B (en) | Full text retrieval inquiry index method for extensible markup language document in relational database | |
| US6122644A (en) | System for halloween protection in a database system | |
| CN103823865A (en) | Database primary memory indexing method | |
| CN102541757B (en) | Write cache method, cache synchronization method and device | |
| CN103164490B (en) | A kind of efficient storage implementation method of not fixed-length data and device | |
| US20020143743A1 (en) | Methods for in-place online reorganization of a database | |
| EP3238106A1 (en) | Compaction policy | |
| CN102890722A (en) | Indexing method applied to time sequence historical database | |
| CN103473298B (en) | Data archiving method and device and storage system | |
| CN102725752A (en) | Method and device for processing dirty data | |
| CN104572920A (en) | Data arrangement method and data arrangement device | |
| WO2005041059A1 (en) | Sorting result buffer | |
| CN103246549A (en) | Method and system for data transfer | |
| CN101408882B (en) | Method and system for searching authorization document | |
| WO2006095356A1 (en) | A method of logging transactions and a method of reversing a transaction | |
| CN101963993B (en) | Method for fast searching database sheet table record | |
| CN109213898A (en) | The video retrieval method and device of video monitoring system | |
| CN103778219B (en) | HBase-based method for updating incremental indexes | |
| CN102768672A (en) | Disk space management method and device | |
| CN108804571B (en) | Data storage method, device and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120905 |