Hibernate的優(yōu)化策略

1、hibernate的優(yōu)化策略hibernate performance tuning (reprint, author: robbin fan)a. inverse =?inverse = false (the default)for oneway one-to-many associationsparentgetchildren ()add (child) / / insert childparent. getchildren ()delete (child) / / delete childinverse = truefor a two-way one-to-many associatio

2、nchild. setparent (parent); the session. save (child) / / insert ch訂dthe session. the delete (child)in a hierarchical structureparentdao, childdao,s encapsulation of crud leads to the ability to persist objects directly through the session interface, but rarely through association objects2. one -to

3、- many relationshipsa oneway relationship or a twoway relationship?parentgetch訂dren ().add (child) the touch action of the collection causes lazy collection initialization, which should be avoided without a secondary cache of the collectionselect * from child where parent_id 二 xxx;the performance fo

4、rmula:in general, avoid one-way connections and try to use a two-way connectionusing a two-way correlation, inverse 二 trueusing the dao interface to persist objects directly through the dao interface in the hierarchical structure, avoiding the ability to persist through association relationships3. m

5、any -to - one relationshipone-way many-to-one expresses foreign key storagethe flexible use of the many-to-one can avoid some unnecessary performance problemsthe meaning of the many-to-one expression is: 0. n: 1, many can be 0, can be 1, or n, which means that many - to one can express one-to-many,

6、one-to-one, many-to-one relationshiptherefore, a two-way many-to-one relationship can be configured, for example:what is the relationship between a table of four players playing mahjong, mahjong seats and mah-jongg people? it's a two-way many-to-one relationshipfour. one -to - oneit is associate

7、d with the primary keythis is the equivalent of breaking up large tables into smaller tablesfor example, break up large fields separately to improve performance of database operationshibernate，s one-to-one seems unable to be lazy and must be passed bytecode enhancementfive. set the list/bag/setone -

8、to - manythe list needs to maintain the index column, can，t be used for bi-directional associations, must be inverse 二"false", and be carefully used in some rare occasionsthere is no difference between a bag/seti personally prefer to use bagmany -to - manythere is a difference between bag

9、and set semantics2. it is recommended to use the setsix. a collection of filtering1 children 二 session. createfilter (parent getchildren (),z/where this .age > 5 and this .age < 10")a very large case for a one-to-many set element, especially for a large collection of pages:().getchildren

10、()setfirstresult (0). setmaxresults (10).in hibernate, supergetsession ()createfilter ()seven. the implicit polymorphism in the inheritance relationshiphql: from the objectcheck out all database tablespolymorphism 二"default" takes the current object, and all the inherited subclasses of the

11、 object, all at oncepolymorphism 二"explicit", which only takes out the current query objecteight hibernate second level cachethe famous n + 1 problem: from child, and then displaying the parent information of each subclass on the page, wi11 result in an n bar query to the parent table:sele

12、ct * from parent where id 二？select * from parent where id 二？the solution1. the eager fetchsecondary cachenine.the relationship between the inverse and the secondary cachewhen using the collection cache:the hibernate maintains the collection cache by using parent getchildren ()inverse 二 /true，/, whic

13、h directly operates on the child and fails to maintain the collection cache! results in cache dirty datathe collection cache should be avoided in the case of a two-way correlation, inverse 二 trueten. the hibernate secondary cache is a magic weapon for improving web application performanceoltp types

14、of web applications, because the application server side can be clustered horizontally, the ultimate system bottleneck always runs away from the database access;which framework minimizes database access, reduces database access pressure, and which framework provides more performance; caching strateg

15、ies for databases:object cache: fine granularity, in view of the level of table records, transparent access, under the condition of without changing the program code can greatly improve the performance of web applications object caching is the magic weapon of orm.the quality of the object cache depe

16、nds on the level of implementation of the framework, which is the most powerful open source orm for the known object cachequery cache: coarse-grained, applied to a situation where data is not high in real-time, for query result sets11. the application context determines the system architecturedo you

17、 need ormhibernate or ibatis?the adoption of orm determines the database designhibernate:tend to fine granularity design, object-oriented, the big table is split into multiple correlation of the small table, eliminate redundant column, through the second level cache performance (dba relationship a t

18、aboo, but the cache orm will break through the correlation performance bottlenecks); the performance bottleneck for hibernate is not the correlation, but the operation of the large tableibatis:tends to the coarse-grained design, to the relationship, as far as possible to merge the table, through the

19、 redundant of the table column, to eliminate the correlation there is no efficient caching method. the performance bottleneck for ibatis is not the large table operations, but the correlation.conclusion:the performance of the formulause a two-way one-to-many relation, not using one direction one-to-

20、manyflexible use of one direction is more relevantyou don，t have to go one-on-one, you have to replace itconfigure the object cache without the collection cachea pair of multiple sets uses bag, with a set for multiple setsthe inheritance class uses explicit polymorphismthere are fewer table fields,

21、and don，t worry about the list correlation. there is a second level cacheyou，ve recently begun to look at the performance issues of hibernate in your project, hoping to find time to learn about the hiberante performance optimization it is mainly about database connection pool technology, hibernate s

22、econdary cache, hibernate configuration optimization, etcrelationships:common correlation: not including a join table, which is the middle table such as:the create table person (personld bigint not null primary key, addressld bigint not null)create table address (addressld bigint not null primary key)