CN103577480B - A kind of parameter division system and its method, a kind of transaction processing system and its method - Google Patents
A kind of parameter division system and its method, a kind of transaction processing system and its method Download PDFInfo
- Publication number
- CN103577480B CN103577480B CN201210277733.0A CN201210277733A CN103577480B CN 103577480 B CN103577480 B CN 103577480B CN 201210277733 A CN201210277733 A CN 201210277733A CN 103577480 B CN103577480 B CN 103577480B
- Authority
- CN
- China
- Prior art keywords
- parameter
- frequent
- read
- caching
- systematic
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2453—Query optimisation
Abstract
The present invention provides a kind of parameter division system and its method, a kind of transaction processing system and its method.The parameter division system includes production environment database, frequent parameter learning module, frequent parameter database, parameter loading module and system parameter data storehouse, the production environment database purchase has same day business datum, the system parameter data stock contains the systematic parameter of flood tide, same day business datum of the wherein described frequent parameter learning module in the Production database generates frequent parameter and by frequent parameter database described in the frequent parameter read-in, the parameter loading module is according to the system parameter data storehouse and the frequent parameter database generation Primary parameter caching and secondary parameters caching.Using the present invention, parameter can be divided, it can in addition contain improve search efficiency.
Description
Technical field
The invention belongs to data processing field, more particularly to a kind of parameter division system and its method and a kind of Business Processing
System and method.
Background technology
With developing rapidly for Bank Card Industry, portfolio is increasing, and batch transaction processing system will often be processed
Hundred million magnanimity transaction data, flood tide business/technical parameter that covering All Activity situation is accessed in processing procedure is right to complete
The accurate treatment of each transaction.As class of business increases, various parameters quantity may proceed to increase, and system can be more and more slower.
So, at present for systematic entirety energy, the search procedure of flood tide parameter has been a very important ring.
At present, typically have following several for the searching method of flood tide parameter:
1st, all parameter full doses load shared drive, and HASH algorithms are realized in shared drive.The effect brought is:
1) search hit rate is very high;
2)Rely on HASH algorithms necessarily require all parameters do not have service priority, without general parameter define so that need
Show configuration it is all determine dimensions all reference records, ginseng high number, it is necessary to shared memory space it is extremely many;
3) workload of parameter configuration personnel is very big, if business rule changes, it is necessary to the parameter note changed and increase newly
Record number is unimaginable, and system the method larger for parameter amount is unrealistic.
2nd, parameter configuration includes business generic definition, loads shared drive, is graded algorithm using class two in internal memory.Bring
Effect be:
1) the search hit rate every time to full dose parameter sets in internal memory is not high, and this depends on fixed comprising business wildcard
The number of the parameter row of justice.
2)Compared with the 1st kind of method, parameter amount is relatively small, and space hold is relatively fewer;
3)Mitigate the workload of parameter configuration personnel.Because thousands of before a business wildcard parameter may be replaced are non-
Business wildcard parameter.
3rd, the high-performance searching algorithm of other some independent researches.The effect brought is:
1)Some current high-performance searching algorithms may bring raising of the search performance on certain proportion;
2)These algorithms often all rely on the business description of parameter, and business rule changes, and algorithm needs adjustment;
3) high-performance algorithm generally requires larger space expense.
The content of the invention
In view of this, the present invention provides a kind of parameter division system and its method and a kind of transaction processing system and its side
Method, is used to divide parameter, improves search efficiency.
The present invention provides following technical scheme:
1. a kind of parameter division system, it is characterised in that including production environment database, frequent parameter learning module, frequency
Numerous parameter database, parameter loading module and system parameter data storehouse, the production environment database purchase have same day business number
According to, the system parameter data stock contains the systematic parameter of flood tide, wherein
Same day business datum of the frequent parameter learning module in the Production database generates frequent parameter simultaneously
By frequent parameter database described in the frequent parameter read-in, the parameter loading module according to the system parameter data storehouse and
The frequent parameter database generation Primary parameter caching and secondary parameters are cached, described in the Primary parameter buffer memory frequently
The frequent parameter that the match is successful in the systematic parameter in parameter, in systematic parameter described in the secondary parameters buffer memory
Unsuccessful systematic parameter is matched in the Primary parameter caching.
2. the system as described in technical scheme 1, it is characterised in that the frequent parameter learning module is in the frequent ginseng of generation
Frequent parameter database is emptied before number and/or after generation Primary parameter caching and secondary parameters caching.
3. the system as described in technical scheme 1 or 2, it is characterised in that the frequent parameter learning module is according to described
Same day business datum in Production database is configured as performing following steps when generating frequent parameter:
(A1)Read the same day business datum in the Production database;
(A2)The each parameter access amount of business datum of statistics;
(A3)The initial frequently parameter sets for the treatment of generation are carried out to statistics using cluster training and hill-climbing algorithm;
(A4)By it is initial frequently in parameter sets the parameter comprising business asterisk wildcard replace with corresponding non-wildcard parameter with
Generate frequent parameter.
4. the system as described in technical scheme 3, it is characterised in that step A3 includes:
(A31)The parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter sets of generation,
Wherein the initial value of K is 2;
(A32)The heuristic function of all business datum searching times of whole day is defined according to currently used searching algorithm, it is right
The K sub- parameter sets use heuristic function;
(A33)K values Jia 1, and the parameter access amount for being counted is clustered using K mean cluster algorithm, generate K son ginseng
Manifold is closed, and [1, K-1] subparameter is merged into initial frequently parameter sets;
(A34)Heuristic function is used to the K sub- parameter sets;
(A35)Adjacent K is carried out using hill-climbing algorithm to current heuristic function result and previous heuristic function result
It is worth the comparing of cluster result searching times, if the searching times after currently the searching times of cluster than clustering before are small, returns
Step A33;Otherwise perform subsequent step.
5. the system as described in one of technical scheme 1-4, it is characterised in that the parameter loading module is in generation one-level
Parameter cache and secondary parameters caching are configured as performing following steps:
(B1)Read all systematic parameters in system parameter data storehouse;
(B2)Read a frequent parameter in frequent parameter database;
(B3)Whether the match is successful in systematic parameter for the frequent parameter that determination is read, if it succeeds, will be read
Frequent parameter read-in Primary parameter caching, otherwise perform step B4;
(B4)Whether the frequent parameter that determination is read is the frequent parameter of the last item, if YES, then performs follow-up step
Suddenly, otherwise return to step B2;
(B5)Read a systematic parameter in system parameter data storehouse;
(B6)Whether the match is successful in Primary parameter caching to determine read systematic parameter, if it fails, then will
The systematic parameter write-in secondary parameters caching for being read, otherwise performs step B7;
(B7)Determine whether read systematic parameter is the last item systematic parameter in system parameter data storehouse, if
It is yes, then exits, otherwise return to step B5.
6. a kind of parameter division methods, it is characterised in that methods described includes:
(A)Frequent parameter is generated and by the frequent frequent parameter database of parameter read-in according to same day business datum;
(B)The frequency in systematic parameter set in advance and the frequent parameter database in system parameter data storehouse
Numerous parameter generation Primary parameter caching and secondary parameters caching, wherein frequent supplemental characteristic described in the Primary parameter buffer memory
The frequent parameter that the match is successful in the systematic parameter in storehouse, system parameter data storehouse described in the secondary parameters buffer memory
In the Primary parameter caching in match unsuccessful systematic parameter.
7. the method as described in technical scheme 6, it is characterised in that methods described also includes:
Frequent parameter is emptied before the frequent parameter of generation and/or after generation Primary parameter caching and secondary parameters caching
Database.
8. the method as described in technical scheme 6 or 7, it is characterised in that step A includes:
(A1)Read the same day business datum in the Production database;
(A2)The each parameter access amount of business datum of statistics;
(A3)The initial frequently parameter sets for the treatment of generation are carried out to statistics using cluster training and hill-climbing algorithm;
(A4)By it is initial frequently in parameter sets the parameter comprising business asterisk wildcard replace with corresponding non-wildcard parameter with
Generate frequent parameter.
9. the method as described in technical scheme 8, it is characterised in that step A3 includes:
(A31)The parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter sets of generation,
Wherein K initial values are 2;
(A32)The heuristic function of all business datum searching times of whole day is defined according to currently used searching algorithm, it is right
The K sub- parameter sets use heuristic function;
(A33)K values Jia 1, and the parameter access amount for being counted is clustered using K mean cluster algorithm, generate K son ginseng
Manifold is closed, and [1, K-1] subparameter is merged into initial frequently parameter sets;
(A34)Heuristic function is used to the K sub- parameter sets;
(A35)Adjacent K is carried out using hill-climbing algorithm to current heuristic function result and previous heuristic function result
It is worth the comparing of cluster result searching times, if the searching times after currently the searching times of cluster than clustering before are small, returns
Step A33;Otherwise perform subsequent step.
10. the method as described in one of technical scheme 6-9, it is characterised in that step B includes:
(B1)Read all systematic parameters in system parameter data storehouse;
(B2)Read a frequent parameter in frequent parameter database;
(B3)Whether the match is successful in systematic parameter for the frequent parameter that determination is read, if it succeeds, will be read
Frequent parameter read-in Primary parameter caching, otherwise perform step B4;
(B4)Whether the frequent parameter that determination is read is the frequent parameter of the last item, if YES, then performs follow-up step
Suddenly, otherwise return to step B2;
(B5)Read a systematic parameter in system parameter data storehouse;
(B6)Whether the match is successful in Primary parameter caching to determine read systematic parameter, if it fails, then will
The systematic parameter write-in secondary parameters caching for being read, otherwise performs step B7;
(B7)Determine whether read systematic parameter is the last item systematic parameter in system parameter data storehouse, if
It is yes, then exits, otherwise return to step B5.
11. a kind of transaction processing systems, it is characterised in that divide system including the parameter as described in one of technical scheme 1-5
System and production batch processing system, wherein the production batch processing system is when business data processing is carried out next day, first in one-level ginseng
Scanned in number caching, if search failure is just scanned for into secondary parameters caching.
12. a kind of method for processing business, it is characterised in that methods described includes:
Enter line parameter using the method as described in one of technical scheme 6-10 to divide;
When next day business data processing is carried out, first scanned in Primary parameter caching, if search failure is just entered
Enter secondary parameters caching to scan for.
Using the present invention, due to being scanned in the Primary parameter caching comprising the frequent parameter for occurring first, so
Improve search efficiency.Secondly, the present invention can follow transaction business rule of development change system adaptively changing Primary parameter
Caching and secondary parameters caching.Do not change current search algorithm, but can guarantee that current search algorithm hit rate highest.It is hardly new
Increase any space expense.Support that parameter configuration librarian use business wildcard parameter is configured, parameter amount is relatively fewer, and space is opened
Pin is small.In addition, the present invention can also ensure the transaction business species that batch processing system more frequently occurs in being concluded the business to magnanimity
Searching times are few, and ensure the overall search least number of times of full dose transaction on the basis of current search strategy, while also dropping
The low workload of parameter configuration personnel.
By test, on the premise of same trading volume, same searching algorithm, the system and conventional solution pair are used
Than the search performance of multiclass parameter is significantly improved in system, according to the cluster difference of the current transaction business rule of development
Lifting ratio is in 30%-80%.The program is a kind of intelligentized parameter training scheme, can be with market business regular shape
Development and change Active Learning, update frequently/non-frequent parameters knowledge, without manual intervention.
Brief description of the drawings
Fig. 1 is the structural representation according to parameter division system of the invention;
Fig. 2 is the schematic flow sheet according to parameter division methods of the invention;
Fig. 3 is the structural representation according to transaction processing system of the invention;And
Fig. 4 is the schematic flow sheet according to parameter division methods of the invention.
Specific embodiment
The preferred embodiments of the present invention are described in detail below in conjunction with accompanying drawing, identical reference number represents phase in the accompanying drawings
Same element.
The present invention is trained based on machine learning-cluster, is a kind of present situation different for all kinds of business development speed, make be
System is according to conventional business development rule(Often 2:8 laws)/ non-frequently internal memory is carried out frequently to current business/technical parameter
Two grades of divisions, and two grades of division results of application memory are substantially improved system in the future to the search hit rate of flood tide parameter.
Fig. 1 is the structural representation according to parameter division system of the invention.As shown in figure 1, parameter division system includes
Production environment database 1, frequent parameter learning module 2, frequent parameter database 3 and parameter loading module 4 and systematic parameter number
According to storehouse 5, the production environment database 4 is stored with same day business datum, and system parameter data storehouse 5 is stored with the system ginseng of flood tide
Number.
Same day business datum of the frequent parameter learning module 2 in Production database 1 generates frequent parameter and by frequently
The frequent parameter database 3 of parameter read-in, parameter loading module 4 is according to system parameter data storehouse 5 set in advance and frequent parameter
The generation Primary parameter of database 3 caching 41 and secondary parameters caching 42, in system in the frequent parameter of storage of Primary parameter caching 41
The frequent parameter that the match is successful in parameter, secondary parameters can not matched in caching 42 storage system parameters in Primary parameter caching
The systematic parameter of work(.
Preferably, frequent parameter learning module 2 is before frequent parameter is generated and/or generation Primary parameter caches 41 and two
Frequent parameter database is emptied after level parameter cache 42.
Preferably, frequent parameter learning module 2 generates frequent parameter in the same day business datum in Production database 1
When be configured as perform following steps:
(A1)Read the same day business datum in the Production database;
(A2)The each parameter access amount of business datum of statistics;
(A3)The initial frequently parameter sets for the treatment of generation are carried out to statistics using cluster training and hill-climbing algorithm to utilize
Cluster training and hill-climbing algorithm are trained generation training result to statistics;
(A4)Parameter comprising business asterisk wildcard in the initial frequently parameter sets of training result is replaced with corresponding non-through
With parameter generating frequent parameter.
Specifically, step A3 is further included:
(A31)The parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter sets of generation,
Wherein the initial value of K is 2;
(A32)The heuristic function of all business datum searching times of whole day is defined according to currently used searching algorithm, it is right
The K sub- parameter sets use heuristic function;
(A33)K values Jia 1, and the parameter access amount for being counted is clustered using K mean cluster algorithm, generate K son ginseng
Manifold is closed, and [1, K-1] subparameter is merged into initial frequently parameter sets;
(A34)Heuristic function is used to the K sub- parameter sets;
(A35)Adjacent K is carried out using hill-climbing algorithm to current heuristic function result and previous heuristic function result
It is worth the comparing of cluster result searching times, if the searching times after currently the searching times of cluster than clustering before are small, returns
Step A33;Otherwise perform subsequent step.
Preferably, parameter loading module 4 is configured as performing in generation Primary parameter caching 41 and the caching of secondary parameters 42
Following steps:
(B1)Read all systematic parameters in system parameter data storehouse;
(B2)Read a frequent parameter in frequent parameter database;
(B3)Whether the match is successful in systematic parameter for the frequent parameter that determination is read, if it succeeds, will be read
Frequent parameter read-in Primary parameter caching, otherwise perform step B4;
(B4)Whether the frequent parameter that determination is read is the frequent parameter of the last item, if YES, then performs follow-up step
Suddenly, otherwise return to step B2;
(B5)Read a systematic parameter in system parameter data storehouse;
(B6)Whether the match is successful in Primary parameter caching to determine read systematic parameter, if it fails, then will
The systematic parameter write-in secondary parameters caching for being read, otherwise performs step B7;
(B7)Determine whether read systematic parameter is the last item systematic parameter in system parameter data storehouse, if
It is yes, then exits, otherwise return to step B5.
Fig. 2 is the schematic flow sheet according to parameter division methods of the invention.As illustrated, the method includes:
(A)Frequent parameter is generated and by the frequent frequent parameter database of parameter read-in according to same day business datum;
(B)The frequency in systematic parameter set in advance and the frequent parameter database in system parameter data storehouse
Numerous parameter generation Primary parameter caching and secondary parameters caching, wherein frequent supplemental characteristic described in the Primary parameter buffer memory
The frequent parameter that the match is successful in the set of system parameters in storehouse, systematic parameter number described in the secondary parameters buffer memory
In gathering according to storehouse unsuccessful systematic parameter is matched in Primary parameter caching.
Preferably, the method also includes:
Frequent parameter is emptied before the frequent parameter of generation and/or after generation Primary parameter caching and secondary parameters caching
Database.
Preferably, step A includes:
(A1)Read the same day business datum in the Production database;
(A2)The each parameter access amount of business datum of statistics;
(A3)The initial frequently parameter sets for the treatment of generation are carried out to statistics using cluster training and hill-climbing algorithm;
(A4)By it is initial frequently in parameter sets the parameter comprising business asterisk wildcard replace with corresponding non-wildcard parameter with
Generate frequent parameter.
Specifically, step A3 includes:
(A31)The parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter sets of generation,
Wherein K initial values are 2;
(A32)The heuristic function of all business datum searching times of whole day is defined according to currently used searching algorithm, it is right
The K sub- parameter sets use heuristic function;
(A33)K values Jia 1, and the parameter access amount for being counted is clustered using K mean cluster algorithm, generate K son ginseng
Manifold is closed, and [1, K-1] subparameter is merged into initial frequently parameter sets;
(A34)Heuristic function is used to the K sub- parameter sets;
(A35)Adjacent K is carried out using hill-climbing algorithm to current heuristic function result and previous heuristic function result
It is worth the comparing of cluster result searching times, if the searching times after currently the searching times of cluster than clustering before are small, returns
Step A33;Otherwise perform subsequent step.
Preferably, step B includes:
(B1)Read all set of system parameters in system parameter data storehouse;
(B2)Read a frequent parameter in frequent parameter database;
(B3)Whether the match is successful in set of system parameters for the frequent parameter that is read of determination, if it succeeds, by institute
The frequent parameter read-in Primary parameter caching for reading, otherwise performs step B4;
(B4)Whether the frequent parameter that determination is read is the frequent parameter of the last item, if YES, then performs follow-up step
Suddenly, otherwise return to step B2;
(B5)Read a systematic parameter in set of system parameters database;
(B6)Whether the match is successful in Primary parameter is cached to determine the parameter in read set of system parameters, if
Systematic parameter write-in secondary parameters caching that is unsuccessful, then will being read, otherwise performs step B7;
(B7)Determine whether read systematic parameter is the last item systematic parameter in set of system parameters database,
If YES, then exit, otherwise return to step B5.
Fig. 3 is the structural representation according to transaction processing system of the invention.As illustrated, the transaction processing system includes
Parameter division system and production batch processing system 6 described in Fig. 1, wherein the production batch processing system 6 carries out business in next day
During data processing, first scanned in Primary parameter caching 41, if search failure is just carried out into secondary parameters caching 42
Search
Fig. 4 is the schematic flow sheet according to parameter division methods of the invention.As illustrated, the method for processing business bag
Include:
In step 401, enter line parameter using the method as described in weight graph 2 and divide.
In step 402, parameter search is carried out.Specifically, when next day business data processing is carried out, first in Primary parameter
Scanned in caching, if search failure is just scanned for into secondary parameters caching.
Using the present invention, transaction business rule of development change system adaptively changing firsts and seconds parameter can be followed to delay
Deposit.Do not change current search algorithm, but can guarantee that current search algorithm hit rate highest.Hardly increase any space expense newly.
In addition, the present invention can also ensure the transaction business species search time that batch processing system more frequently occurs in being concluded the business to magnanimity
Number is few, and ensures the overall search least number of times of full dose transaction on the basis of current search strategy, while also reducing ginseng
The workload of number configuration personnel.
In view of these instruct, those of ordinary skill in the art will readily occur to other embodiments of the invention, combination and
Modification.Therefore, when described above is combined and accompanying drawing is read, the present invention is only defined by the claims.
Claims (10)
1. a kind of parameter division system, it is characterised in that including production environment database, frequently frequent parameter learning module, ginseng
Number database, parameter loading module and system parameter data storehouse, the production environment database purchase have same day business datum, institute
The systematic parameter that system parameter data stock contains flood tide is stated, wherein
Same day business datum of the frequent parameter learning module in the production environment database generates frequent parameter simultaneously
By frequent parameter database described in the frequent parameter read-in, the parameter loading module according to the system parameter data storehouse and
The frequent parameter database generation Primary parameter caching and secondary parameters are cached, described in the Primary parameter buffer memory frequently
The frequent parameter that the match is successful in the systematic parameter in parameter, in systematic parameter described in the secondary parameters buffer memory
Unsuccessful systematic parameter is matched in the Primary parameter caching,
Wherein, the parameter loading module is configured as performing following step in generation Primary parameter caching and secondary parameters caching
Suddenly:
(B1)Read all systematic parameters in system parameter data storehouse;
(B2)Read a frequent parameter in frequent parameter database;
(B3)Whether the match is successful in systematic parameter for the frequent parameter that determination is read, if it succeeds, the frequency that will be read
Numerous parameter read-in Primary parameter caching, otherwise performs step B4;
(B4)Whether the frequent parameter that determination is read is the frequent parameter of the last item, if YES, then performs subsequent step, no
Then return to step B2;
(B5)Read a systematic parameter in system parameter data storehouse;
(B6)Whether the match is successful in Primary parameter caching to determine read systematic parameter, if it fails, will then be read
The systematic parameter write-in secondary parameters caching for taking, otherwise performs step B7;
(B7)Determine whether read systematic parameter is the last item systematic parameter in system parameter data storehouse, if
It is then to exit, otherwise return to step B5.
2. the system as claimed in claim 1, it is characterised in that the frequent parameter learning module is before frequent parameter is generated
And/or empty frequent parameter database after generation Primary parameter caching and secondary parameters caching.
3. system as claimed in claim 1 or 2, it is characterised in that the frequent parameter learning module is according to the production
Same day business datum in environment data base is configured as performing following steps when generating frequent parameter:
(A1)Read the same day business datum in the production environment database;
(A2)The each parameter access amount of business datum of statistics;
(A3)The initial frequently parameter sets for the treatment of generation are carried out to statistics using cluster training and hill-climbing algorithm;
(A4)Parameter comprising business asterisk wildcard in initial frequently parameter sets is replaced with into corresponding non-wildcard parameter to generate
Frequent parameter.
4. system as claimed in claim 3, it is characterised in that step A3 includes:
(A31)The parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter sets of generation, wherein
The initial value of K is 2;
(A32)The heuristic function of all business datum searching times of whole day is defined according to currently used searching algorithm, to the K
Subparameter set uses heuristic function;
(A33)K values Jia 1, and the parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter set of generation
Close, [1, K-1] subparameter is merged into initial frequently parameter sets;
(A34)Heuristic function is used to the K sub- parameter sets;
(A35)Adjacent K values are carried out to current heuristic function result and previous heuristic function result using hill-climbing algorithm to gather
The comparing of class result searching times, if the searching times after currently the searching times of cluster than clustering before are small, return to step
A33;Otherwise perform subsequent step.
5. a kind of parameter division methods, it is characterised in that methods described includes:
(A)Frequent parameter is generated and by the frequent frequent parameter database of parameter read-in according to same day business datum;
(B)The frequent ginseng in systematic parameter set in advance and the frequent parameter database in system parameter data storehouse
Number generation Primary parameter caching and secondary parameters caching, wherein in frequent parameter database described in the Primary parameter buffer memory
The frequent parameter that the match is successful in the systematic parameter, in system parameter data storehouse described in the secondary parameters buffer memory
Unsuccessful systematic parameter is matched in the Primary parameter caching,
Wherein, step B includes:
(B1)Read all systematic parameters in system parameter data storehouse;
(B2)Read a frequent parameter in frequent parameter database;
(B3)Whether the match is successful in systematic parameter for the frequent parameter that determination is read, if it succeeds, the frequency that will be read
Numerous parameter read-in Primary parameter caching, otherwise performs step B4;
(B4)Whether the frequent parameter that determination is read is the frequent parameter of the last item, if YES, then performs subsequent step, no
Then return to step B2;
(B5)Read a systematic parameter in system parameter data storehouse;
(B6)Whether the match is successful in Primary parameter caching to determine read systematic parameter, if it fails, will then be read
The systematic parameter write-in secondary parameters caching for taking, otherwise performs step B7;
(B7)Determine whether read systematic parameter is the last item systematic parameter in system parameter data storehouse, if
It is then to exit, otherwise return to step B5.
6. method as claimed in claim 5, it is characterised in that methods described also includes:
Frequent supplemental characteristic is emptied before the frequent parameter of generation and/or after generation Primary parameter caching and secondary parameters caching
Storehouse.
7. the method as described in claim 5 or 6, it is characterised in that step A includes:
(A1)Read the same day business datum in the production environment database;
(A2)The each parameter access amount of business datum of statistics;
(A3)The initial frequently parameter sets for the treatment of generation are carried out to statistics using cluster training and hill-climbing algorithm;
(A4)Parameter comprising business asterisk wildcard in initial frequently parameter sets is replaced with into corresponding non-wildcard parameter to generate
Frequent parameter.
8. method as claimed in claim 7, it is characterised in that step A3 includes:
(A31)The parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter sets of generation, wherein
K initial values are 2;
(A32)The heuristic function of all business datum searching times of whole day is defined according to currently used searching algorithm, to the K
Subparameter set uses heuristic function;
(A33)K values Jia 1, and the parameter access amount for being counted is clustered using K mean cluster algorithm, K sub- parameter set of generation
Close, [1, K-1] subparameter is merged into initial frequently parameter sets;
(A34)Heuristic function is used to the K sub- parameter sets;
(A35)Adjacent K values are carried out to current heuristic function result and previous heuristic function result using hill-climbing algorithm to gather
The comparing of class result searching times, if the searching times after currently the searching times of cluster than clustering before are small, return to step
A33;Otherwise perform subsequent step.
9. a kind of transaction processing system, it is characterised in that including the parameter division system as described in one of claim 1-4 and life
Batch processing system is produced, wherein the production batch processing system is when business data processing is carried out next day, first in Primary parameter caching
In scan for, if search failure just into secondary parameters caching scan for.
10. a kind of method for processing business, it is characterised in that methods described includes:
Enter line parameter using the method as described in one of claim 5-8 to divide;
When next day business data processing is carried out, first scanned in Primary parameter caching, if search failure just enters two
Level parameter cache is scanned for.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210277733.0A CN103577480B (en) | 2012-08-07 | 2012-08-07 | A kind of parameter division system and its method, a kind of transaction processing system and its method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210277733.0A CN103577480B (en) | 2012-08-07 | 2012-08-07 | A kind of parameter division system and its method, a kind of transaction processing system and its method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103577480A CN103577480A (en) | 2014-02-12 |
CN103577480B true CN103577480B (en) | 2017-05-31 |
Family
ID=50049278
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210277733.0A Active CN103577480B (en) | 2012-08-07 | 2012-08-07 | A kind of parameter division system and its method, a kind of transaction processing system and its method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103577480B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109933614B (en) * | 2019-02-01 | 2021-08-06 | 青岛微智慧信息有限公司 | Electronic tax bureau data preprocessing method based on big data behavior analysis |
CN112529169A (en) * | 2019-09-18 | 2021-03-19 | 华为技术有限公司 | Data processing method, model optimization device and model execution device |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1295687A (en) * | 1998-03-31 | 2001-05-16 | 英特尔公司 | Shared cache structure for temporal and non-temporal instructions |
CN1606006A (en) * | 2004-10-27 | 2005-04-13 | 浪潮电子信息产业股份有限公司 | Record section storing method for database management system |
CN101216855A (en) * | 2008-01-16 | 2008-07-09 | 中兴通讯股份有限公司 | Message kind business accumulated data caching method and system |
CN101576918A (en) * | 2009-06-19 | 2009-11-11 | 用友软件股份有限公司 | Data buffering system with load balancing function |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7321954B2 (en) * | 2004-08-11 | 2008-01-22 | International Business Machines Corporation | Method for software controllable dynamically lockable cache line replacement system |
US7793049B2 (en) * | 2007-10-30 | 2010-09-07 | International Business Machines Corporation | Mechanism for data cache replacement based on region policies |
-
2012
- 2012-08-07 CN CN201210277733.0A patent/CN103577480B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1295687A (en) * | 1998-03-31 | 2001-05-16 | 英特尔公司 | Shared cache structure for temporal and non-temporal instructions |
CN1606006A (en) * | 2004-10-27 | 2005-04-13 | 浪潮电子信息产业股份有限公司 | Record section storing method for database management system |
CN101216855A (en) * | 2008-01-16 | 2008-07-09 | 中兴通讯股份有限公司 | Message kind business accumulated data caching method and system |
CN101576918A (en) * | 2009-06-19 | 2009-11-11 | 用友软件股份有限公司 | Data buffering system with load balancing function |
Non-Patent Citations (1)
Title |
---|
一种基于网格和密度凝聚点的快速聚类算法;陈卓等;《哈尔滨工业大学学报》;20051230;第1655页第2.1节第1-2段 * |
Also Published As
Publication number | Publication date |
---|---|
CN103577480A (en) | 2014-02-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220292103A1 (en) | Information service for facts extracted from differing sources on a wide area network | |
CN106777351B (en) | Computing system and its method are stored based on ART tree distributed system figure | |
US9892150B2 (en) | Unified data management for database systems | |
CN103593436B (en) | file merging method and device | |
Huang et al. | Naive Bayes classification algorithm based on small sample set | |
CN104462362B (en) | A kind of data storage, inquiry, loading method and device | |
CN103186566B (en) | A kind of data classification storage, apparatus and system | |
CN103631922B (en) | Extensive Web information extracting method and system based on Hadoop clusters | |
CN102521406A (en) | Distributed query method and system for complex task of querying massive structured data | |
CN102402627B (en) | System and method for real-time intelligent capturing of article | |
CN109034194A (en) | Transaction swindling behavior depth detection method based on feature differentiation | |
CN101593202A (en) | Based on the hash connecting method for database of sharing the Cache polycaryon processor | |
CN109086412A (en) | A kind of unbalanced data classification method based on adaptive weighted Bagging-GBDT | |
US20210149923A1 (en) | Systems and methods for intelligently grouping financial product users into cohesive cohorts | |
CN107784105A (en) | Construction of knowledge base method, electronic installation and storage medium based on magnanimity problem | |
CN109558421A (en) | Data processing method, system, device and storage medium based on caching | |
CN104407879A (en) | A power grid timing sequence large data parallel loading method | |
CN110990529B (en) | Industry detail dividing method and system for enterprises | |
CN106339331A (en) | Data cache layer replacement algorithm based on user activity degree | |
Huang et al. | An efficient automatic multiple objectives optimization feature selection strategy for internet text classification | |
CN103577480B (en) | A kind of parameter division system and its method, a kind of transaction processing system and its method | |
CN101739429A (en) | Method for optimizing cluster search results and device thereof | |
CN108052535A (en) | The parallel fast matching method of visual signature and system based on multi processor platform | |
Kirshners et al. | A comparative analysis of short time series processing methods | |
CN113127567A (en) | Financial data computing system, method and device based on distributed storage |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |