2. 程式人生 > >Hibernate快取原理與查詢快取的組合探究

Hibernate快取原理與查詢快取的組合探究

阿新 • • 發佈：2019-02-08

來源：http://www.javaeye.com/topic/431603

0.前言

         由於對Hibernate的二級快取和查詢快取的區別不瞭解，也不知道它們起什麼作用。於是動手做了一些實驗，對它們的組合使用有了一個表面的認識。

1.前提

         1) 不使用一級快取（Session級別）的情況。因為大部分api對session的操作都進行n層封裝，用完session就close，一般很少使用到一級快取。

         2) 開啟hibernate.show_sql=true，根據sql的是否輸出來判斷是否訪問了資料庫。

2.基礎

2.1.二級快取

         二級快取是SessionFactory級別的全域性快取，它為每個類（或集合）提供快取。凡是呼叫二級快取的查詢方法都會從中受益，比如load,list,iterate等方法（注意，get和find不使用快取，直接訪問資料庫）。

         開啟條件：

         1)hibernate.cache.provider_class=org.hibernate.cache.EhCacheProvider（這裡指定了ehcache）

         2)hbm.xml配置<cache usage="read-write"/>

         3)ehcache.xml定義該類的cache

         簡單地說，二級快取是用來儲存類物件的，並且填充它的所有屬性值。

2.2.查詢快取

         查詢快取儲存的是查詢方法（list和iterate等）查詢的sql語句和結果集，跟load這種裝載物件的方法沒有關係。也就是說，使用load方法，查詢快取不理睬。

         並且這個結果集只是物件id列表，物件的其他屬性值不儲存。例如，list執行的時候，先根據查詢快取中的sql語句，獲取物件id列表，然後再load(id)來取物件。

         開啟條件：

         1) hibernate.cache.use_query_cache=true

         2)Query或者Criteria使用時，設定setCacheable(true)

         簡單地說，查詢快取是用來儲存sql執行後獲取的物件id列表。下一次執行同一條sql時，可直接從查詢快取中獲取到物件id列表。

3.實驗

         以下對最常用的load和list方法進行二級快取和查詢快取的組合測試。load和list執行的結果，都是返回同一個物件。

3.1.四種組合

         二級快取和查詢快取的組合，根據開啟和不開啟兩種情況，劃分為四種組合（√代表開啟，×代表不開啟）：

組合

二級快取

查詢快取

1

√

×

2

×

√

3

√

√

4

×

×

3.2.load實驗

         對同一個id，使用load執行兩次，觀察sql的輸出現象，也就是判斷第二次load是否訪問了資料庫。其中√代表輸出，×代表未輸出：

組合

第一次sql

第二次sql

結論

1

√

×

與查詢快取無關

2

√

√

3

√

×

4

√

√

         當開啟了二級快取，load先從快取中獲取物件，於是組合1和3的第二次sql都未輸出，即第二次查詢沒有訪問資料庫。

         當沒有開啟二級快取，load總是直接訪問資料庫的。也證明了，查詢快取與load方法無關。

3.3.list實驗

         對同一條sql語句，使用list執行兩次，觀察sql的輸出現象。其中√代表輸出，×代表未輸出：

組合

第一次sql

第二次sql

結論

1

√

√

list每次訪問資料庫都會填充二級快取，相當於二級快取無效

2

√

√（不一樣）

這兩條sql語句不一樣，第一條是list的，第二條是load的

3

√

×

4

√

√

         對組合1現象的解釋：

         當沒有開啟查詢快取時，list每次都會直接訪問資料庫，然後把二級快取給重新填充。也就是說，list在不開啟查詢快取的情況下，根本就沒有利用到二級快取的好處。

對組合2現象的解釋：

         第一條sql是list正常訪問資料庫的，但第二次執行list時，由於有了查詢快取，直接根據第一條sql獲取物件的id，此時由於沒有開啟二級快取，load(id)後也沒法從二級快取中拿資料，只好再次訪問資料庫。但在組合3中，load(id)可以從二級快取中拿資料，所以不用訪問資料庫。

3.4.小結

         在load和list方法中，使用二級快取和查詢快取的不同組合，效果出現了差別。特別強調的是，由於很多操作是基於list的，如果沒有開啟查詢快取，根本就使用不了二級快取的資料，也就提高不了效能了（但是iterate可以使用到二級快取）。

         另外，從表格中可以看出來，同時開啟二級快取和查詢快取，效果是最好的，第二次查詢都不需要訪問資料庫。

4.維護

         當存在除hibernate之外的改變資料庫資料的方式時，維護二級快取和查詢快取是很麻煩的。比如使用jdbc或者第三方系統修改了資料庫，hibernate都不知道，也就沒法自行維護快取了。這時需要手動維護，如果存在第三方系統，還需要讓第三方系統發訊息通知hibernate。

         維護手段也就是evict（清除某個類的二級快取）和evictQuery（清除查詢快取）。針對資料修改的三種形式，分別維護：

         1)insert：evictQuery

         2)delete：evict和evictQuery

         3)update：evict

一般來講ORM中的快取分為以下幾類:
         1:事務級快取:即在當前事務範圍內的資料快取. 就Hibernate來講,事務級快取是基於Session的生命週期實現的,每個Session內部會存在一個數據快取,它隨著 Session的建立而存在,隨著Session的銷燬而滅亡,因此也稱為Session Level Cache.
         2:應用級快取: 即在某個應用中或應用中某個獨立資料庫訪問子集中的共享快取,此快取可由多個事務共享(資料庫事務或應用事務),事務之間的快取共享策略與應用的事務隔離 機制密切相關.在Hibernate中,應用級快取由SessionFactory實現,所有由一個SessionFactory建立的 Session例項共享此快取,因此也稱為SessionFactory Level Cache.
         3:分散式快取:即在多個應用例項,多個JVM間共享的快取策略.分散式快取由多個應用級快取例項組成,通過某種遠端機制(RMI,JMS)實現各個快取例項間的資料同步,任何一個例項的資料修改,將導致整個叢集間的資料狀態同步.

  Hibernate的二級快取策略，是針對於ID查詢的快取策略，對於條件查詢則毫無作用。為此，Hibernate提供了針對條件查詢的Query Cache。

查詢快取，快取的key是hql語句，快取的value是滿足hql語句的記錄的主鍵值。也就是說，查詢快取，只是快取資料庫記錄的主鍵值，並不會快取記錄的所有欄位值。

    預設hibernate不會開啟查詢快取，這是因為查詢快取只有在hql/hql語句語義完全一致的時候，才能命中。而實際查詢場景下，查詢條件、分頁、 排序等構成的複雜查詢sql語句很難完全一致。可能是hibernate覺得命中率低，所以預設關閉了查詢快取。我們可以根據實際使用情況，決定是否開啟 查詢快取，唯一的原則就是命中率要儘可能的高。如果針對A表的查詢，查詢sql語句基本都是完全一致的情況，就可以針對A使用查詢快取；如果B表的查詢條 件經常變化，很難命中，那麼就不要對B表使用查詢快取。這可能就是hibernate使用查詢快取的時候，既要在hibernate.cfg.xml中進 行配置，也需要query.setCacheable(true)的原因。

     查詢快取只對list有用，對iterate方式無用。iterate不會讀也不會寫查詢快取，list會讀也會寫查詢快取。查詢快取中的key是 sql語句（這些sql語句會被hibernate解析，保證語義相同的sql，能夠命中查詢快取），快取的value是記錄的主鍵值

Q:什麼樣的資料適合存放到第二級快取中？

    1.很少被修改的資料

    2.不是很重要的資料，允許出現偶爾併發的資料

    3.不會被併發訪問的資料

    4.參考資料,指的是供應用參考的常量資料，它的例項數目有限，它的例項會被許多其他類的例項引用，例項極少或者從來不會被修改。

　不適合存放到第二級快取的資料？

    1 經常被修改的資料

    2 財務資料，絕對不允許出現併發

    3 與其他應用共享的資料。

 hibernate快取的資料實時性思考？當資料變化時hibernate快取是否能自動更新？

下面是測試例子，經過測試hibernate查詢快取能“監聽”到當前快取實體的資料變化，即當對這個實體進行增加，刪除，修改時，查詢快取將重新去資料庫獲得最新的資料然後再快取起來。

前置配置：

http://www.360doc.com/content/14/0801/16/1073512_398635409.shtml

注意在實體對映檔案中配置：<!--二級快取一般設定為只讀--><cache usage="read-only" />

實驗：返回的是list

測試程式碼：

 DAO層的一個方法：

 public void testSession(){
        Session session=this.getSession();
    //List<TUser>    ls=session.createQuery("from TUser where username like '%h%'").setFirstResult(0).setMaxResults(2).list();
      List<TUser>    ls=session.createQuery("from TUser where username like ? ").setCacheable(true).setParameter(0, "%h%").setFirstResult(0).setMaxResults(30).list();
        for (TUser tUser : ls) {
            System.out.println(tUser.getUsername()+"=="+tUser.getId());
        }
    }

junit單元測試的一個方法：

    @Test
    public void testCache() {
        
        System.out.println("=====第一次讀取=======");
        userManager.testCache();
        System.out.println("====第二次讀取========");
        userManager.testCache();
        System.out.println("====新增一個使用者=======");
        TUser user=new TUser();
        user.setUsername("htt");
        user.setPassword("123");
        userManager.saveUser(user);
        System.out.println("====新增使用者後第一次讀取======");
        userManager.testCache();
        System.out.println("====新增使用者後第二次讀取======");
        userManager.testCache();
        System.out.println("====修改使用者=====");
        TUser user2=new TUser();
        user2.setId("1061");
        user2.setUsername("h22");
        user2.setPassword("123");
        userManager.update(user2);
        System.out.println("====修改使用者後第一次讀取======");
        userManager.testCache();
        System.out.println("====修改使用者後第二次讀取======");
        userManager.testCache();
        System.out.println("====刪除使用者=====");
        TUser user3=new TUser();
        user3.setId("1061");
        userManager.deleteUser(user3);
        System.out.println("====刪除使用者後第一次讀取======");
        userManager.testCache();
        System.out.println("====刪除使用者後第二次讀取======");
        userManager.testCache();
    }

場景一：開啟“查詢快取”，關閉二級快取。結果如下

=====第一次讀取=======
Hibernate: select * from ( select tuser0_.ID as ID1_, tuser0_.USERNAME as USERNAME1_, tuser0_.PASSWORD as PASSWORD1_, tuser0_.EMAIL as EMAIL1_, tuser0_.MOBILE as MOBILE1_ from hzc.T_USER tuser0_ where tuser0_.USERNAME like ? ) where rownum <= ?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
h22==1061
htt==1082
htt==1083
htt==1085
====第二次讀取========
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
h22==1061
htt==1082
htt==1083
htt==1085
====新增一個使用者=======
Hibernate: select SEQ_USER.nextval from dual
Hibernate: insert into hzc.T_USER (USERNAME, PASSWORD, EMAIL, MOBILE, ID) values (?, ?, ?, ?, ?)
====新增使用者後第一次讀取======
Hibernate: select * from ( select tuser0_.ID as ID1_, tuser0_.USERNAME as USERNAME1_, tuser0_.PASSWORD as PASSWORD1_, tuser0_.EMAIL as EMAIL1_, tuser0_.MOBILE as MOBILE1_ from hzc.T_USER tuser0_ where tuser0_.USERNAME like ? ) where rownum <= ?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
h22==1061
htt==1082
htt==1083
htt==1085
htt==1088
====新增使用者後第二次讀取======
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
h22==1061
htt==1082
htt==1083
htt==1085
htt==1088
====修改使用者=====
Hibernate: update hzc.T_USER set USERNAME=?, PASSWORD=?, EMAIL=?, MOBILE=? where ID=?
====修改使用者後第一次讀取======
Hibernate: select * from ( select tuser0_.ID as ID1_, tuser0_.USERNAME as USERNAME1_, tuser0_.PASSWORD as PASSWORD1_, tuser0_.EMAIL as EMAIL1_, tuser0_.MOBILE as MOBILE1_ from hzc.T_USER tuser0_ where tuser0_.USERNAME like ? ) where rownum <= ?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
h22==1061
htt==1082
htt==1083
htt==1085
htt==1088
====修改使用者後第二次讀取======
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
h22==1061
htt==1082
htt==1083
htt==1085
htt==1088
====刪除使用者=====
Hibernate: delete from hzc.T_USER where ID=?
====刪除使用者後第一次讀取======
Hibernate: select * from ( select tuser0_.ID as ID1_, tuser0_.USERNAME as USERNAME1_, tuser0_.PASSWORD as PASSWORD1_, tuser0_.EMAIL as EMAIL1_, tuser0_.MOBILE as MOBILE1_ from hzc.T_USER tuser0_ where tuser0_.USERNAME like ? ) where rownum <= ?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
htt==1082
htt==1083
htt==1085
htt==1088
====刪除使用者後第二次讀取======
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
Hibernate: select tuser0_.ID as ID1_0_, tuser0_.USERNAME as USERNAME1_0_, tuser0_.PASSWORD as PASSWORD1_0_, tuser0_.EMAIL as EMAIL1_0_, tuser0_.MOBILE as MOBILE1_0_ from hzc.T_USER tuser0_ where tuser0_.ID=?
htt==1081
htt==1087
htt==1086
htt==1084
hcc==1010
hzc==1004
htt==1082
htt==1083
htt==1085
htt==1088

場景一：開啟“查詢快取”，開啟二級快取。結果如下

log4j:WARN No appenders could be found for logger (org.springframework.test.context.junit4.SpringJUnit4ClassRunner).
log4j:WARN Please initialize the log4j system properly.
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
=====第一次讀取=======
Hibernate: select * from ( select tuser0_.ID as ID1_, tuser0_.USERNAME as USERNAME1_, tuser0_.PASSWORD as PASSWORD1_, tuser0_.EMAIL as EMAIL1_, tuser0_.MOBILE as MOBILE1_ from hzc.T_USER tuser0_ where tuser0_.USERNAME like ? ) where rownum <= ?
htt==1081
htt==1087
htt==1086
htt==1084
htt==1089
hcc==1010
hzc==1004
htt==1082
htt==1083
htt==1085
htt==1088
====第二次讀取========
htt==1081
htt==1087
htt==1086
htt==1084
htt==1089
hcc==1010
hzc==1004
htt==1082
htt==1083
htt==1085
htt==1088
====新增一個使用者=======
Hibernate: select SEQ_USER.nextval from dual
Hibernate: insert into hzc.T_USER (USERNAME, PASSWORD, EMAIL, MOBILE, ID) values (?, ?, ?, ?, ?)
====新增使用者後第一次讀取======
Hibernate: select * from ( select tuser0_.ID as ID1_, tuser0_.USERNAME as USERNAME1_, tuser0_.PASSWORD as PASSWORD1_, tuser0_.EMAIL as EMAIL1_, tuser0_.MOBILE as MOBILE1_ from hzc.T_USER tuser0_ where tuser0_.USERNAME like ? ) where rownum <= ?
htt==1081
htt==1087
htt==1086
htt==1084
htt==1089
hcc==1010
hzc==1004
htt==1082
htt==1083
htt==1085
htt==1088
htt==1090
====新增使用者後第二次讀取======
htt==1081
htt==1087
htt==1086
htt==1084
htt==1089
hcc==1010
hzc==1004
htt==1082
htt==1083
htt==1085
htt==1088
htt==1090
====修改使用者=====


相關推薦

Hibernate快取原理與查詢快取的組合探究

來源：http://www.javaeye.com/topic/431603 0.前言          由於對Hibernate的二級快取和查詢快取的區別不瞭解，也不知道它們起什麼作用。於是動手做了一些實驗，對它們的組合使用有了一個表面的認識。 1.前提      

Hibernate快取原理與策略

Hibernate快取原理      對於Hibernate這類ORM而言,快取顯的尤為重要,它是持久層效能提升的關鍵.簡單來講Hibernate就是對JDBC進行封裝,以實現內部狀態的管理,OR關係的對映等,但隨之帶來的就是資料訪問效率的降低,和效能的下降,而快取就是彌

hibernate二級快取與查詢快取之間的關係

我們先看一下什麼樣的hibernate操作會將資料存入到一級快取中去，當呼叫下邊這些方法的時候會將資料存入到一級快取中 一級快取 save、update、saveOrupdate、load、list、iterate、lock，例如： User userOne=new Use

Hibernate快取體系之查詢快取（query cache），以及list和iterate方法的區別

       Hibernate框架提供了Session.load()和Session.get()方法，用來根據實體物件的主鍵值從資料庫中查詢對應記錄。針對load和get方法，hibernate提供了一級快取和二級快取的支援，提高查詢效率，具體可以參考我的部落格：通過測試

hibernate快取之【查詢快取】

        說到hibernate中的快取我們瞭解的比較多的是一級和二級快取，通過前面兩篇部落格的總結我們知道一級和二級快取主要是快取查詢的實體或實體集的，那麼如果我們是通過HQL或SQL語句查詢

Integer快取原理與JVM調參應用

基本資料型別的拆箱與裝箱 jvm把一個包裝型別自動轉成基本資料型別叫做拆箱，反之叫做裝箱。例如： Integer integer = 2 ;//裝箱 實際執行程式碼為Integer integer = Integer.valueOf(2); int

mysql 開啟查詢快取方法與查詢例子

開啟快取，設定快取大小，具體實施如下： 1、修改配置檔案,windows下是my.ini，linux下是my.cnf; 在配置檔案的最後追加上：  程式碼如下 複製程式碼 query_cache_type = 1 query_cache_size = 600000 需要

hibernate物件狀態與一級快取二級快取

1 hibernate 中 持久類物件狀態介紹 1.1hibernate 規定三種狀態：瞬時態、持久態、脫管態      瞬時態：transient，     session沒有快取物件，資料庫也沒有對應記錄。         OID特點：沒有值     持久態：per

快取原理與微服務快取自動管理

> 拋開業務談技術都是在耍流氓。—— Kevin Wan ## 為什麼需要快取？ 先從一個老生常談的問題開始談起：我們的程式是如何執行起來的？ 1. 程式儲存在 `disk` 中 2. 程式是執行在 `RAM` 之中，也就是我們所說的 `main memory` 3. 程式的計算邏輯在 `CPU` 中執

瀏覽器快取cookie與伺服器快取

之前對快取都是一知半解的，都是基於理論之上沒有正整實際動手驗證過，今天抽時間終於把快取重新理了一片 1、 設定js 、css緩存       http頭部中設定expires,Cache-Contro

Android Glide快取清除與獲取快取大小

不多說了，還是直接上程式碼吧 /** * Created by zhaoyong on 2016/6/21. * Glide快取工具類 */ public class ImageCatchUtil { private static ImageCatchUtil inst; priva

iOS 獲取快取大小與清除快取

首先要獲取cache資料夾路徑 #define cachePath [NSSearchPathForDirectoriesInDomains(NSCachesDirectory, NSUserDomainMask, YES) firstObject] //

MySQL索引的最左字首原理與查詢的相關優化

最基礎：索引的使用，儘量能夠做到一一對應每條資料表中的資料。 1）全列匹配，當按照索引中所有列進行精確匹配時，索引可以被用到，理論上，索引 對順序是敏感的，但是由於MySQL的查詢優化器會自動調整where子句中 的條件順序以使用合適的索引。 2)最左字首匹配，當查詢條件

Android Glide獲取快取大小與清除快取

GlideCatchSimple SimpleDemo請看github Glide快取Simple 快取路徑的指定 快取大小的獲取 磁碟快取清除(兩種方法) 記憶體快取清除 可clone之後檢視使用Simple Glide cache Si

hibernate 二級快取和查詢快取原理和關係

一、hibernate的二級快取 如果開啟了二級快取，hibernate在執行任何一次查詢的之後，都會把得到的結果集放到快取中，快取結構可以看作是一個hash table，key是資料庫記錄的id，value是id對應的pojo物件。當用戶根據id查詢物件的時候（load、iterator方法），會首先在

Mybatis(3、延遲載入、查詢快取、與ehcache整合、逆向工程、與spring整合)

版權宣告：本文為博主原創文章，未經博主允許不得轉載。    https://blog.csdn.net/www1056481167/article/details/70597788 延遲載入 延遲載入：先從單表查詢、需要時再從關聯表去關聯查詢，大大提高 資料庫效能，因為

Hibernate 二級快取和查詢快取

一級快取:     1,在session上面有一個一級快取;一級快取的生命週期和session相同,一級快取最大生命週期就是一個執行緒;在web環境下面,session的最大生命週期就是一次請求;     2,一級快取可以用來幹嘛? &nb

mybatis基礎系列（四）——關聯查詢、延遲載入、一級快取與二級快取

關本文是Mybatis基礎系列的第四篇文章，點選下面連結可以檢視前面的文章： mybatis基礎系列（三）——動態sql mybatis基礎系列（二）——基礎語法、別名、輸入對映、輸出對映 mybatis基礎系列（一）——mybatis入門 關聯查詢 在進行表設計時，往往需要在具體的業務基礎上分析表與表之間的

Hibernate 髒資料檢查機制與資料快取

一. 髒資料檢查機制 髒資料:   髒資料並非廢棄或者無用的資料,而是指一個數據物件所攜帶的資訊發生了改變之後的狀態.            如果我們從資料庫中讀取一個物件

Hibernate的一級快取、二級快取和查詢快取

Hibernate的Session提供了一級快取的功能，預設總是有效的，當應用程式儲存持久化實體、修改持久化實體時，Session並不會立即把這種改變提交到資料庫，而是快取在當前的Session中，除非顯示呼叫了Session的flush()方法或通過close()方法關閉Sessi