如何加快SQL查詢

Question

我需要建議是否可以進行此調整。

表：

CREATE TABLE UTIL_DAY (
  CompID varchar(15) collate utf8_czech_ci NOT NULL,
  TestplanID int(10) unsigned NOT NULL default '0',
  FromDate datetime NOT NULL default '0000-00-00 00:00:00',
  ToDate datetime NOT NULL default '0000-00-00 00:00:00',
  Interval int(11) NOT NULL default '0',
  Variant varchar(25) collate utf8_czech_ci default NULL,
  WO_ID int(11) default NULL,
  SAPTime int(5) default NULL,
  UseTime int(5) NOT NULL,
  MeasureCount int(10) unsigned NOT NULL,
  KEY CompID (CompID),
  KEY TestplanID (TestplanID),
  KEY WO_ID (WO_ID),
  KEY FromDate (FromDate),
  KEY Variant (Variant)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_czech_ci;

當前查詢：

UPDATE UTIL_DAY UDAY SET TestplanID = (SELECT TestplanID 
                                       FROM (SELECT UD_U.Variant,UD_U.WO_ID,VS.TestplanID 
                                             FROM (SELECT UD.Variant,UD.WO_ID 
                                                   FROM UTIL_DAY UD WHERE SUBSTR(CompID,1,3) <> 'AOI' 
                                                   GROUP BY UD.Variant,UD.WO_ID) UD_U 
                                             JOIN VARIANTS V ON UD_U.Variant = V.Variant 
                                             JOIN VARIANT_STATUS VS ON V.VariantID = VS.VariantID 
                                             JOIN TESTPLANS_WO_ID TW ON UD_U.WO_ID = TW.WO_ID 
                                             WHERE VS.TestplanID = TW.TestplanID 
                                             GROUP BY UD_U.Variant,UD_U.WO_ID ) TP 
                                       WHERE TP.Variant = UDAY.Variant AND TP.WO_ID = UDAY.WO_ID)

我的更新是與臨時表：

CREATE TEMPORARY TABLE UTIL_DAY_TEMP (
Variant varchar(25) COLLATE 'utf8_czech_ci' NOT NULL,
WO_ID int unsigned NOT NULL,
TestplanID int unsigned NOT NULL,
PRIMARY KEY Variant_WO_ID (Variant, WO_ID)) COMMENT='' ENGINE='InnoDB' COLLATE 'utf8_czech_ci';
INSERT IGNORE INTO UTIL_DAY_TMP (SELECT UD_U.Variant,UD_U.WO_ID,VS.TestplanID 
                FROM (SELECT UD.Variant,UD.WO_ID 
            FROM UTIL_DAY UD WHERE SUBSTR(CompID,1,3) <> 'AOI' 
            GROUP BY UD.Variant,UD.WO_ID) UD_U 
      JOIN VARIANTS V ON UD_U.Variant = V.Variant 
      JOIN VARIANT_STATUS VS ON V.VariantID = VS.VariantID 
      JOIN TESTPLANS_WO_ID TW ON UD_U.WO_ID = TW.WO_ID 
      WHERE VS.TestplanID = TW.TestplanID 
      GROUP BY UD_U.Variant,UD_U.WO_ID);
 UPDATE UTIL_DAY UDAY SET TestplanID = (SELECT TestplanID 
                                        FROM UTIL_DAY_TEMP TP 
                                         WHERE TP.Variant = UDAY.Variant AND TP.WO_ID = UDAY.WO_ID);

此改進使對200萬行的查詢速度提高了400％。 但是這種改進有相同的結果嗎？ 謝謝您的回答。

Answer 1

首先，使用

來自mysql的EXPLAIN命令。

之后，找出是否有任何索引對您有幫助（警告：不要在無法控制的情況下創建索引）。此外，請查看mysql上復合索引的文檔