Is there a way to increase performance of a query which uses multiple subquerys in select clause
我有一个查询，它从两个表 TABLE_1 和 TABLE_2 中选择数据。这两个表是用id连接的。 TABLE_1 中的每一行在 TABLE_2 上可以正好有 4 行。客户需要一个视图，该视图返回 TABLE_1 中的所有数据和 TABLE_2 中的所有数据，但 TABLE_2 中的数据必须显示为与同一行中的列一样。例如：
这是我用过的一个查询：
1 2 3 4 5 6 7 8 9 | SELECT ID, NAME, LASTNAME, (SELECT HomeType FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 0 ROW FETCH NEXT 1 ROW ONLY) AS Row1Column1, (SELECT HomeCost FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 0 ROW FETCH NEXT 1 ROW ONLY) AS Row1Column2, (SELECT HomeType FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 1 ROW FETCH NEXT 1 ROW ONLY) AS Row2Column1, (SELECT HomeCost FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 1 ROW FETCH NEXT 1 ROW ONLY) AS Row2Column2, (SELECT HomeType FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 2 ROW FETCH NEXT 1 ROW ONLY) AS Row2Column1, (SELECT HomeCost FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 2 ROW FETCH NEXT 1 ROW ONLY) AS Row2Column2, (SELECT HomeType FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 3 ROW FETCH NEXT 1 ROW ONLY) AS Row2Column1, (SELECT HomeCost FROM #TABLE_2 WHERE id = t1.ID ORDER BY ID OFFSET 3 ROW FETCH NEXT 1 ROW ONLY) AS Row2Column2 FROM #TABLE_1 AS t1 |
我的数据的一个例子是
1 2 3 4 5 6 7 8 9 10 | CREATE TABLE #TABLE_1(ID INT, NAME VARCHAR(100), LASTNAME VARCHAR(100)) CREATE TABLE #TABLE_2(ID INT,HomeType VARCHAR(100), HomeCost VARCHAR(100) ) INSERT INTO #TABLE_1 (ID, NAME, LASTNAME) VALUES (1, 'JOHN', 'SNOW') INSERT INTO #TABLE_2 (ID, HomeType, HomeCost) VALUES (1, 'Type1', 'Cost1') INSERT INTO #TABLE_2 (ID, HomeType, HomeCost) VALUES (1, 'Type2', 'Cost2') INSERT INTO #TABLE_2 (ID, HomeType, HomeCost) VALUES (1, 'Type3', 'Cost3') INSERT INTO #TABLE_2 (ID, HomeType, HomeCost) VALUES (1, 'Type4', 'Cost4') |
我想要这样的结果，但是对于 50000 行它会变慢
1 2 3 | ID| NAME| ROW1COLUMN1 | ROW1COLUMN2 | ROW2COLUMN1 | ROW2COLUMN2 | ROW3COLUMN1 | ROW3COLUMN2 | ROW4COLUMN1 | ROW4COLUMN2 ------------------------- 1|JOHN | Type1 | Cost1 | Type2 | Cost2 | Type3 | Cost3 | Type4 | Cost4 |
我没有测试这个查询，但它应该可以工作。不要忘记为 ID、HomeType 字段做索引。
如果字段"HomeType"是动态的，您需要添加额外的字段"pos"。对于 TABLE_2 中具有相同 ID 的行，"pos"字段应该是唯一的。
1 2 3 4 5 6 | SELECT t.ID, t.NAME, t.LASTNAME, t1.HomeType AS Row1Column1, t1.HomeCost AS Row1Column2, t2.HomeType AS Row2Column1, t2.HomeCost AS Row2Column2, t3.HomeType AS Row3Column1, t3.HomeCost AS Row3Column2, t4.HomeType AS Row4Column1, t4.HomeCost AS Row4Column2 FROM TABLE_1 t JOIN TABLE_2 t1 ON t.ID = t1.ID AND t1.HomeType = 'Type1' JOIN TABLE_2 t2 ON t.ID = t2.ID AND t2.HomeType = 'Type2' JOIN TABLE_2 t3 ON t.ID = t3.ID AND t3.HomeType = 'Type3' JOIN TABLE_2 t4 ON t.ID = t4.ID AND t4.HomeType = 'Type4' |
您可以使用 PIVOT
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 | WITH T2 AS ( SELECT ID, HomeTypeRN = 'A' + CAST( ROW_NUMBER() OVER (PARTITION BY ID ORDER BY HomeType) AS VARCHAR), HomeType, HomeCostRN = 'B' + CAST(ROW_NUMBER() OVER (PARTITION BY ID ORDER BY HomeType) AS VARCHAR), HomeCost FROM #TABLE_2 ) SELECT T1.ID, T1.Name, T1.LastName, [A1], [B1], [A2], [B2], [A3], [B3], [A4], [B4] FROM #TABLE_1 T1 LEFT JOIN ( SELECT ID, MAX(A1) AS A1, MAX(B1) AS B1, MAX(A2) AS A2, MAX(B2) AS B2, MAX(A3) AS A3, MAX(B3) AS B3, MAX(A4) AS A4, MAX(B4) AS B4 FROM T2 PIVOT( MAX(HomeType) FOR HomeTypeRN IN ([A1],[A2],[A3],[A4]) ) P1 PIVOT( MAX(HomeCost) FOR HomeCostRN IN ([B1],[B2],[B3],[B4]) ) P2 GROUP BY ID ) P ON P.ID = T1.ID |
如果你有相同的四种类型，那么你可以像这样使用条件聚合：
1 2 3 4 5 6 7 8 9 10 11 12 13 | SELECT t1.id, t1.name, t1.lastname, MAX(CASE WHEN t2.type = 'type1' THEN hometype END) AS hometype_1, MAX(CASE WHEN t2.type = 'type1' THEN homecost END) AS homecost_1, MAX(CASE WHEN t2.type = 'type2' THEN hometype END) AS hometype_2, MAX(CASE WHEN t2.type = 'type2' THEN homecost END) AS homecost_2, MAX(CASE WHEN t2.type = 'type3' THEN hometype END) AS hometype_3, MAX(CASE WHEN t2.type = 'type3' THEN homecost END) AS homecost_3, MAX(CASE WHEN t2.type = 'type4' THEN hometype END) AS hometype_4, MAX(CASE WHEN t2.type = 'type4' THEN homecost END) AS homecost_4 FROM table1 t1 LEFT JOIN t2 ON t1.id = t2.id GROUP BY t1.id, t1.name, t1.lastname; |