[20201224]sql優化困惑.txt
[20201224]sql優化困惑.txt
--//昨天優化生產系統一條語句，不小心折騰一個下午，回家的路上才想起如何優化，自己一下子沒有轉過來，浪費大量的時間。
1.環境：
> @ ver1
PORT_STRING VERSION BANNER
------------------------------ -------------- --------------------------------------------------------------------------------
x86_64/Linux 2.4.xx 11.2.0.4.0 Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
> @ bind_cap 1n7yhk3p9cd66 ''
C200
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
select * from ( select huanzhexin0_.id as col_0_0_, huanzhexin0_.binglihao as col_1_0_, huanzhexin0_.xingming as col_2_0_, huanzhexin0_.xingbie as col_3_0_, huanzhexin0_.jtdz as col_4_0_, huanzhexin0_
.shengri as col_5_0_, jcd2_.id as col_6_0_, jiuzhen1_.id as col_7_0_, jcd2_.jcdh as col_8_0_, jcd2_.kd_time as col_9_0_, jcd2_.kdys as col_10_0_, (select yuangong3_.xingming from yuangong yuangong3_ w
here yuangong3_.gonghao=jcd2_.kdys) as col_11_0_, (select bumen4_.bmmc from bumen bumen4_ where bumen4_.id=jcd2_.kdks_id) as col_12_0_, jcd2_.kdks_id as col_13_0_, jcd2_.biaoti as col_14_0_, (select y
uangong5_.xingming from yuangong yuangong5_ where yuangong5_.gonghao=jcd2_.jcys) as col_15_0_, jcd2_.jcjs_time as col_16_0_ from huanzhexinxi huanzhexin0_, jiuzhen jiuzhen1_, jcd jcd2_ where huanzhexi
n0_.id=jiuzhen1_.huanzhe_id and jiuzhen1_.id=jcd2_.jiuzhen_id and jcd2_.biaoshi=:"SYS_B_0" and (jcd2_.state is null) order by jcd2_.jcjs_time ) where rownum <= :1
SQL_ID CHILD_NUMBER WAS NAME POSITION MAX_LENGTH LAST_CAPTURED DATATYPE_STRING VALUE_STRING
------------- ------------ --- ---------------------------------------- ------------ ------------ ------------------- --------------- ------------
1n7yhk3p9cd66 0 YES :SYS_B_0 1 22 2020-07-21 11:26:31 NUMBER 56
YES :1 2 22 2020-07-21 11:26:31 NUMBER 100
--//理論講這樣語句不會輸出許多行，邏輯讀也不會很高。猜測操作完成後修改state非NULL。
> @ d_buffer 1n7yhk3p9cd66 60
EXECUTIONS1 BUFFER_GETS1 ELAPSED_TIME1 ROWS_PROCESSED1 每次buffer_gets 每次執行時間 平均處理記錄數
--------------- --------------- --------------- --------------- --------------- --------------- ---------------
2098552 576708129 360817107935 32134 274.81240827008 171936.2245658 .01531246306977
... sleep 60 , waiting ....
EXECUTIONS2 BUFFER_GETS2 ELAPSED_TIME2 ROWS_PROCESSED2 每次buffer_gets 每次執行時間 平均處理記錄數
--------------- --------------- --------------- --------------- --------------- --------------- ---------------
2098562 576797449 360819016026 32134 274.85366122135 171936.31449821 .01531239010332
總buffer_gets 每次buffer_gets 執行次數 總執行時間 每次執行時間 總處理記錄數 平均處理記錄數
--------------- --------------- --------------- --------------- --------------- --------------- ---------------
89320 8932 10 1908091 190809.1 0 0
--//每分鐘執行10次。又是一個經典的重新整理語句。每次接近0.19秒。而且經常是輸出記錄為0.
--//執行計劃如下：
Plan hash value: 1469246793
----------------------------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | E-Rows |E-Bytes|E-Temp | Cost (%CPU)| E-Time | OMem | 1Mem | Used-Mem |
----------------------------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | | 4736 (100)| | | | |
| 1 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 1 | 13 | | 2 (0)| 00:00:01 | | | |
|* 2 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 1 | | | 1 (0)| 00:00:01 | 1025K| 1025K| |
| 3 | TABLE ACCESS BY INDEX ROWID | BUMEN | 1 | 14 | | 1 (0)| 00:00:01 | | | |
|* 4 | INDEX UNIQUE SCAN | PK_BUMEN | 1 | | | 0 (0)| | 1025K| 1025K| |
| 5 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 1 | 13 | | 2 (0)| 00:00:01 | | | |
|* 6 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 1 | | | 1 (0)| 00:00:01 | 1025K| 1025K| |
|* 7 | COUNT STOPKEY | | | | | | | | | |
| 8 | VIEW | | 46745 | 107M| | 4736 (1)| 00:00:57 | | | |
|* 9 | SORT ORDER BY STOPKEY | | 46745 | 6893K| 7800K| 4736 (1)| 00:00:57 | 1024 | 1024 | |
|* 10 | HASH JOIN | | 46745 | 6893K| 4848K| 3276 (1)| 00:00:40 | 8577K| 2802K| 9729K (0)|
| 11 | TABLE ACCESS STORAGE FULL FIRST ROWS | HUANZHEXINXI | 75182 | 3964K| | 370 (1)| 00:00:05 | 1025K| 1025K| |
|* 12 | HASH JOIN | | 46745 | 4427K| 3120K| 2430 (1)| 00:00:30 | 9672K| 4837K| 7815K (0)|
| 13 | TABLE ACCESS STORAGE FULL FIRST ROWS| JIUZHEN | 145K| 1416K| | 372 (1)| 00:00:05 | 1025K| 1025K| |
|* 14 | TABLE ACCESS STORAGE FULL FIRST ROWS| JCD | 46745 | 3971K| | 1687 (1)| 00:00:21 | 1025K| 1025K| |
----------------------------------------------------------------------------------------------------------------------------------------------------
2.優化過程略：
--//我加入的提示如下：
/*+ gather_plan_statistics
leading(jcd2_ jiuzhen1_ huanzhexin0_)
index(jcd2_ I_JCD_BIAOSHI_STATE_JCJS_TIME)
index(jiuzhen1_ PK_JIUZHEN)
index(huanzhexin0_ PK_HUANZHEXINXI)
use_nl( jiuzhen1_)
use_nl( huanzhexin0_ )
cardinality(jcd2_ 200)
*/
--// I_JCD_BIAOSHI_STATE_JCJS_TIME 索引包含3個欄位 BIAOSHI，STATE，JCJS_TIME。
--//如果我帶入jcd2_.biaoshi=:"SYS_B_0"的值53，biaoshi=53,state is null很多，執行統計資訊如下：
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows |E-Bytes| Cost (%CPU)| E-Time | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | | | 4214 (100)| | 100 |00:00:00.44 | 164K| | | |
| 1 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 6370 | 1 | 13 | 2 (0)| 00:00:01 | 6354 |00:00:00.02 | 1927 | | | |
|* 2 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 6370 | 1 | | 1 (0)| 00:00:01 | 6354 |00:00:00.01 | 3 | 1025K| 1025K| |
| 3 | TABLE ACCESS BY INDEX ROWID | BUMEN | 6 | 1 | 14 | 1 (0)| 00:00:01 | 6 |00:00:00.01 | 9 | | | |
|* 4 | INDEX UNIQUE SCAN | PK_BUMEN | 6 | 1 | | 0 (0)| | 6 |00:00:00.01 | 3 | 1025K| 1025K| |
| 5 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 15 | 1 | 13 | 2 (0)| 00:00:01 | 14 |00:00:00.01 | 7 | | | |
|* 6 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 15 | 1 | | 1 (0)| 00:00:01 | 14 |00:00:00.01 | 3 | 1025K| 1025K| |
|* 7 | COUNT STOPKEY | | 1 | | | | | 100 |00:00:00.44 | 164K| | | |
| 8 | VIEW | | 1 | 200 | 472K| 4214 (1)| 00:00:51 | 100 |00:00:00.44 | 164K| | | |
|* 9 | SORT ORDER BY STOPKEY | | 1 | 200 | 30400 | 4214 (1)| 00:00:51 | 100 |00:00:00.44 | 164K| 22528 | 22528 |20480 (0)|
| 10 | NESTED LOOPS | | 1 | 200 | 30400 | 4213 (1)| 00:00:51 | 48454 |00:00:00.39 | 162K| | | |
| 11 | NESTED LOOPS | | 1 | 200 | 30400 | 4213 (1)| 00:00:51 | 48454 |00:00:00.31 | 114K| | | |
| 12 | NESTED LOOPS | | 1 | 200 | 19600 | 4013 (1)| 00:00:49 | 48454 |00:00:00.22 | 70013 | | | |
| 13 | TABLE ACCESS BY INDEX ROWID| JCD | 1 | 200 | 17600 | 3813 (1)| 00:00:46 | 48454 |00:00:00.07 | 5407 | | | |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|* 14 | INDEX RANGE SCAN | I_JCD_BIAOSHI_STATE_JCJS_TIME | 1 | 9315 | | 45 (0)| 00:00:01 | 48454 |00:00:00.02 | 97 | 1025K| 1025K| |
| 15 | TABLE ACCESS BY INDEX ROWID| JIUZHEN | 48454 | 1 | 10 | 1 (0)| 00:00:01 | 48454 |00:00:00.12 | 64606 | | | |
|* 16 | INDEX UNIQUE SCAN | PK_JIUZHEN | 48454 | 1 | | 0 (0)| | 48454 |00:00:00.06 | 16152 | 1025K| 1025K| |
|* 17 | INDEX UNIQUE SCAN | PK_HUANZHEXINXI | 48454 | 1 | | 0 (0)| | 48454 |00:00:00.07 | 44430 | 1025K| 1025K| |
| 18 | TABLE ACCESS BY INDEX ROWID | HUANZHEXINXI | 48454 | 1 | 54 | 1 (0)| 00:00:01 | 48454 |00:00:00.06 | 48454 | | | |
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
--//實際上掃描jcd獲得100條記錄就可以停止，而實際讀取48454行，如果控制呢？感覺oracle的優化器那裡出了問題。
3.分析：
--//我在這裡折騰很長時間，一直沒有調整出來。後來我單獨把表jcd拿出來執行，才知道問題在哪裡。
alter session set statistics_level = all;
select * from (select id from jcd where biaoshi=53 and state is null order by jcjs_time) where rownum<=10;
Plan hash value: 1740280897
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows |E-Bytes| Cost (%CPU)| E-Time | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | | | 1188 (100)| | 10 |00:00:00.24 | 1070 | | | |
|* 1 | COUNT STOPKEY | | 1 | | | | | 10 |00:00:00.24 | 1070 | | | |
| 2 | VIEW | | 1 | 8191 | 103K| 1188 (1)| 00:00:15 | 10 |00:00:00.24 | 1070 | | | |
|* 3 | SORT ORDER BY STOPKEY | | 1 | 8191 | 175K| 1188 (1)| 00:00:15 | 10 |00:00:00.24 | 1070 | 2048 | 2048 | 2048 (0)|
|* 4 | VIEW | index$_join$_002 | 1 | 8191 | 175K| 1186 (1)| 00:00:15 | 48496 |00:00:00.22 | 1070 | | | |
|* 5 | HASH JOIN | | 1 | | | | | 48644 |00:00:00.21 | 1070 | 4967K| 3974K| 4037K (0)|
|* 6 | INDEX RANGE SCAN | I_JCD_BIAOSHI_STATE_JCJS_TIME | 1 | 8191 | 175K| 225 (0)| 00:00:03 | 48644 |00:00:00.01 | 98 | 1025K| 1025K| |
| 7 | INDEX STORAGE FAST FULL SCAN| PK_JCD | 1 | 8191 | 175K| 1200 (1)| 00:00:15 | 331K|00:00:00.06 | 972 | 1025K| 1025K| |
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
--//實際讀取48644行。id=6.我希望就是id＝6，僅僅讀取100行就ok了。怎麼控制不住呢。浪費N多時間。
--//回家的路上才想起來，實際上問題出在order by的選擇上。加入biaoshi,state就可以控制僅僅掃描很少的行。
select * from (select id from jcd where biaoshi=53 and state is null order by biaoshi,state,jcjs_time) where rownum<=10;
--//補充一點，只要order by 中jcjs_time不在第一的位置，執行計劃都是一樣的。
Plan hash value: 2856453538
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows |E-Bytes| Cost (%CPU)| E-Time | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | | | 8 (100)| | 10 |00:00:00.01 | 12 | | | |
|* 1 | COUNT STOPKEY | | 1 | | | | | 10 |00:00:00.01 | 12 | | | |
| 2 | VIEW | | 1 | 10 | 130 | 8 (0)| 00:00:01 | 10 |00:00:00.01 | 12 | | | |
| 3 | TABLE ACCESS BY INDEX ROWID| JCD | 1 | 10 | 220 | 8 (0)| 00:00:01 | 10 |00:00:00.01 | 12 | | | |
|* 4 | INDEX RANGE SCAN | I_JCD_BIAOSHI_STATE_JCJS_TIME | 1 | | | 3 (0)| 00:00:01 | 41 |00:00:00.01 | 4 | 1025K| 1025K| |
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
--//這樣就可以很好的控制邏輯讀，掃描行數。
--//這樣上面的語句修改如下：
SELECT *
FROM ( SELECT /*+ gather_plan_statistics */
huanzhexin0_.id AS col_0_0_
,huanzhexin0_.binglihao AS col_1_0_
,huanzhexin0_.xingming AS col_2_0_
,huanzhexin0_.xingbie AS col_3_0_
,huanzhexin0_.jtdz AS col_4_0_
,huanzhexin0_.shengri AS col_5_0_
,jcd2_.id AS col_6_0_
,jiuzhen1_.id AS col_7_0_
,jcd2_.jcdh AS col_8_0_
,jcd2_.kd_time AS col_9_0_
,jcd2_.kdys AS col_10_0_
, (SELECT yuangong3_.xingming
FROM yuangong yuangong3_
WHERE yuangong3_.gonghao = jcd2_.kdys)
AS col_11_0_
, (SELECT bumen4_.bmmc
FROM bumen bumen4_
WHERE bumen4_.id = jcd2_.kdks_id)
AS col_12_0_
,jcd2_.kdks_id AS col_13_0_
,jcd2_.biaoti AS col_14_0_
, (SELECT yuangong5_.xingming
FROM yuangong yuangong5_
WHERE yuangong5_.gonghao = jcd2_.jcys)
AS col_15_0_
,jcd2_.jcjs_time AS col_16_0_
FROM huanzhexinxi huanzhexin0_, jiuzhen jiuzhen1_, jcd jcd2_
WHERE huanzhexin0_.id = jiuzhen1_.huanzhe_id
AND jiuzhen1_.id = jcd2_.jiuzhen_id
AND jcd2_.biaoshi = :"SYS_B_0"
AND (jcd2_.state IS NULL)
ORDER BY jcd2_.biaoshi,jcd2_.state,jcd2_.jcjs_time)
WHERE ROWNUM <= :1;
--//執行計劃如下: :"SYS_B_0" = 53.
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows |E-Bytes| Cost (%CPU)| E-Time | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | | | 32 (100)| | 100 |00:00:00.01 | 353 | | | |
| 1 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 29 | 1 | 13 | 2 (0)| 00:00:01 | 17 |00:00:00.01 | 4 | | | |
|* 2 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 29 | 1 | | 1 (0)| 00:00:01 | 17 |00:00:00.01 | 3 | 1025K| 1025K| |
| 3 | TABLE ACCESS BY INDEX ROWID | BUMEN | 3 | 1 | 14 | 1 (0)| 00:00:01 | 3 |00:00:00.01 | 5 | | | |
|* 4 | INDEX UNIQUE SCAN | PK_BUMEN | 3 | 1 | | 0 (0)| | 3 |00:00:00.01 | 2 | 1025K| 1025K| |
| 5 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 15 | 1 | 13 | 2 (0)| 00:00:01 | 14 |00:00:00.01 | 7 | | | |
|* 6 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 15 | 1 | | 1 (0)| 00:00:01 | 14 |00:00:00.01 | 3 | 1025K| 1025K| |
|* 7 | COUNT STOPKEY | | 1 | | | | | 100 |00:00:00.01 | 353 | | | |
| 8 | VIEW | | 1 | 11 | 26587 | 32 (0)| 00:00:01 | 100 |00:00:00.01 | 353 | | | |
| 9 | NESTED LOOPS | | 1 | 11 | 1672 | 32 (0)| 00:00:01 | 100 |00:00:00.01 | 337 | | | |
| 10 | NESTED LOOPS | | 1 | 12 | 1672 | 32 (0)| 00:00:01 | 100 |00:00:00.01 | 237 | | | |
| 11 | NESTED LOOPS | | 1 | 12 | 1176 | 20 (0)| 00:00:01 | 100 |00:00:00.01 | 196 | | | |
| 12 | TABLE ACCESS BY INDEX ROWID| JCD | 1 | 9315 | 800K| 8 (0)| 00:00:01 | 100 |00:00:00.01 | 46 | | | |
|* 13 | INDEX RANGE SCAN | I_JCD_BIAOSHI_STATE_JCJS_TIME | 1 | 12 | | 3 (0)| 00:00:01 | 100 |00:00:00.01 | 3 | 1025K| 1025K| |
| 14 | TABLE ACCESS BY INDEX ROWID| JIUZHEN | 100 | 1 | 10 | 1 (0)| 00:00:01 | 100 |00:00:00.01 | 150 | | | |
|* 15 | INDEX UNIQUE SCAN | PK_JIUZHEN | 100 | 1 | | 0 (0)| | 100 |00:00:00.01 | 50 | 1025K| 1025K| |
|* 16 | INDEX UNIQUE SCAN | PK_HUANZHEXINXI | 100 | 1 | | 0 (0)| | 100 |00:00:00.01 | 41 | 1025K| 1025K| |
| 17 | TABLE ACCESS BY INDEX ROWID | HUANZHEXINXI | 100 | 1 | 54 | 1 (0)| 00:00:01 | 100 |00:00:00.01 | 100 | | | |
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
--//執行計劃如下: :"SYS_B_0" = 56.這個邏輯讀更少。
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows |E-Bytes| Cost (%CPU)| E-Time | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | | | 32 (100)| | 0 |00:00:00.01 | 3 | | | |
| 1 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 0 | 1 | 13 | 2 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | | | |
|* 2 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 0 | 1 | | 1 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | 1025K| 1025K| |
| 3 | TABLE ACCESS BY INDEX ROWID | BUMEN | 0 | 1 | 14 | 1 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | | | |
|* 4 | INDEX UNIQUE SCAN | PK_BUMEN | 0 | 1 | | 0 (0)| | 0 |00:00:00.01 | 0 | 1025K| 1025K| |
| 5 | TABLE ACCESS BY INDEX ROWID | YUANGONG | 0 | 1 | 13 | 2 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | | | |
|* 6 | INDEX RANGE SCAN | I_YUANGONG_GONGHAO | 0 | 1 | | 1 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | 1025K| 1025K| |
|* 7 | COUNT STOPKEY | | 1 | | | | | 0 |00:00:00.01 | 3 | | | |
| 8 | VIEW | | 1 | 11 | 26587 | 32 (0)| 00:00:01 | 0 |00:00:00.01 | 3 | | | |
| 9 | NESTED LOOPS | | 1 | 11 | 1672 | 32 (0)| 00:00:01 | 0 |00:00:00.01 | 3 | | | |
| 10 | NESTED LOOPS | | 1 | 12 | 1672 | 32 (0)| 00:00:01 | 0 |00:00:00.01 | 3 | | | |
| 11 | NESTED LOOPS | | 1 | 12 | 1176 | 20 (0)| 00:00:01 | 0 |00:00:00.01 | 3 | | | |
| 12 | TABLE ACCESS BY INDEX ROWID| JCD | 1 | 9315 | 800K| 8 (0)| 00:00:01 | 0 |00:00:00.01 | 3 | | | |
|* 13 | INDEX RANGE SCAN | I_JCD_BIAOSHI_STATE_JCJS_TIME | 1 | 12 | | 3 (0)| 00:00:01 | 0 |00:00:00.01 | 3 | 1025K| 1025K| |
| 14 | TABLE ACCESS BY INDEX ROWID| JIUZHEN | 0 | 1 | 10 | 1 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | | | |
|* 15 | INDEX UNIQUE SCAN | PK_JIUZHEN | 0 | 1 | | 0 (0)| | 0 |00:00:00.01 | 0 | 1025K| 1025K| |
|* 16 | INDEX UNIQUE SCAN | PK_HUANZHEXINXI | 0 | 1 | | 0 (0)| | 0 |00:00:00.01 | 0 | 1025K| 1025K| |
| 17 | TABLE ACCESS BY INDEX ROWID | HUANZHEXINXI | 0 | 1 | 54 | 1 (0)| 00:00:01 | 0 |00:00:00.01 | 0 | | | |
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
3.總結：
--//實際上還是細節很重要,如果當時我加入提示實際上優化已經完成，這個語句帶入就是56，根本不會變，記錄很少，邏輯讀不會很高
--//，但是仔細再思考一步，就可以發現開發沒有寫好sql語句。
--//另外再次說明交流很重要，根本沒必要在上面浪費這麼多時間，估計問一下週圍的人許多很快得到結果。可惜在我周圍連一個問問題
--//的人都沒有..........
來自 “ ITPUB部落格 ” ，連結：http://blog.itpub.net/267265/viewspace-2744838/，如需轉載，請註明出處，否則將追究法律責任。
相關文章
- [20200320]SQL語句優化的困惑.txtSQL優化
- [20200324]SQL語句優化的困惑2.txtSQL優化
- [20220428]優化的困惑12.txt優化
- [20210203]max優化的困惑.txt優化
- [20200401]優化的困惑5.txt優化
- [20200408]優化的困惑6.txt優化
- [20220507]優化的困惑13.txt優化
- [20200808]優化的困惑10.txt優化
- [20170104]一條sql優化.txtSQL優化
- [20240313]使用tpt ashtop.sql指令碼的困惑.txtSQL指令碼
- [20131204]sql語句優化.txtSQL優化
- [20151221]sql語句優化.txtSQL優化
- 【SQL優化】SQL優化工具SQL優化
- SQL Server優化之SQL語句優化SQLServer優化
- [20201210]sql語句優化.txtSQL優化
- [20150611]優化sql遇到問題.txt優化SQL
- SQL優化SQL優化
- with as優化sql優化SQL
- 效能優化案例-SQL優化優化SQL
- [20181119]使用sql profile優化問題.txtSQL優化
- 資料庫優化 - SQL優化資料庫優化SQL
- sql優化之邏輯優化SQL優化
- SQL優化：limit分頁優化SQL優化MIT
- MySQL 效能優化之SQL優化MySql優化
- SQL優化--用各種hints優化一條SQLSQL優化
- 【SQL優化】SQL優化的10點注意事項SQL優化
- SQL SERVER中SQL優化SQLServer優化
- SQL優化指南SQL優化
- SQL SERVER優化SQLServer優化
- sql效能優化SQL優化
- Sql優化方法SQL優化
- oracle sql優化OracleSQL優化
- SQL優化-索引SQL優化索引
- SQL優化（一）SQL優化
- oracle sql 優化OracleSQL優化
- sql 效能優化SQL優化
- [20181114]一條sql語句的優化.txtSQL優化
- [20131025]一條sql語句的優化.txtSQL優化