提高内连接的SQL（lite）查询速度

Question

我有以下查询运行很长时间＆gt; 30分钟，我想知道是否有一些方法可以改善性能。

脚本通过Python（3.X）

执行

加入的目标是在表格中填写US_vehicles的车辆详情。

表格设计说明

• FinalUS包含带有单个vin-number（车辆识别号码）的车辆
• US_vehicles包含分配了多个毒蛇的车辆的详细信息（以逗号分隔）

已经采取的措施：

• vin-column（以逗号分隔）被整理成vin1-vin34
• 所有需要匹配的列的索引
• 使用IN代替LIKE

表格信息：

• FinalUS：包含99740行
• US_vehicles：包含62865行

INSERT INTO mergenhtsaed 
            (COMMENT,                 finalusid, 
             edmundsids,              addvin, 
             length,                  height, 
             widthnomirror,           bedlength, 
             bedlengthstd,            doors, 
             torqueprimary,           torqueprimaryrpm, 
             horsepowerrpmprimary,    transtypestd, 
             transspeed,              transtypestddetail, 
             drivenwheels,            cafecarlineclass, 
             wheelbase,               fronttrack, 
             reartrack,               primaryfuelstd, 
             horsepowerprimary,       hybrid, 
             flexfuel,                bodystyle) 
SELECT "1:", 
       FU.id, 
       Group_concat(DISTINCT UV.id), 
       Group_concat(DISTINCT squishvins), 
       Round(Avg(UV.overall_length), 1), 
       Round(Avg(UV.overall_height), 1), 
       Round(Avg(UV.overall_width_without_mirrors), 1), 
       Round(Avg(bedlength), 1), 
       Round(Avg(bedsizestdft), 1), 
       Group_concat(DISTINCT UV.numofdoors), 
       Round(Avg(UV.engine_torque)), 
       Round(Avg(UV.engine_rpm_torque)), 
       Round(Avg(UV.engine_rpm_horsepower)), 
       UV.transtypestd, 
       UV.transspeed, 
       UV.transtypestddetail, 
       UV.drivenwheelsstd, 
       Group_concat(DISTINCT uv .` categories_epaclass ` ), 
       Round(Avg(UV.wheelbase), 1), 
       Round(Avg(UV.front_track), 1), 
       Round(Avg(UV.rear_track), 1), 
       Group_concat(DISTINCT UV.primaryfuelstd), 
       Round(Avg(UV.engine_horsepower)), 
       Group_concat(DISTINCT UV.hybrid), 
       Group_concat(DISTINCT UV.flexfuel), 
       Group_concat(DISTINCT UV.bodystyle) 
FROM   finalus FU 
inner join us_vehicles UV 
        ON UV.years_year = FU.modelyear 
       AND fu .` vin ` IN ( UV.vin1, UV.vin2, UV.vin3, UV.vin4, 
                            UV.vin5, UV.vin6, UV.vin7, UV.vin8, 
                            UV.vin9, UV.vin10, UV.vin11, UV.vin12, 
                            UV.vin13, UV.vin14, UV.vin15, UV.vin16, 
                            UV.vin17, UV.vin18, UV.vin19, UV.vin20, 
                            UV.vin21, UV.vin22, UV.vin23, UV.vin24, 
                            UV.vin25, UV.vin26, UV.vin27, UV.vin28, 
                            UV.vin29, UV.vin30, UV.vin31, UV.vin32, 
                            UV.vin33, UV.vin34 ) 
WHERE FU.modelyear < 2012 
  AND FU.edmundsid IS NULL 
  AND FU.dataoneid IS NULL 
GROUP  BY FU.id 
HAVING Count(DISTINCT UV.id) = 1 

解释查询计划

Selectid|order|from|detail        
"0"|"0"|"0"|"SEARCH TABLE FinalUS AS FU USING INDEX Idx3 (EdmundsId=?)"
"0"|"1"|"1"|"SCAN TABLE US_Vehicles AS UV"