我目前正在编写一个SQL语句,在满足要求时,每个ID只能选择一行。
我假设表COMPONENTS
看起来像这样:
COMP_ID DATE_FROM TIMESTAMP COMP_TRX_ID
1 10-JAN-14 10-JAN-15 5
1 10-JAN-15 10-JAN-14 5
1 10-JAN-12 10-JAN-13 5
1 11-JAN-16 10-JAN-12 5
2 11-JAN-14 10-JAN-15 5
2 11-JAN-15 10-JAN-14 5
2 11-JAN-12 10-JAN-13 5
2 11-JAN-16 10-JAN-12 5
现在让我们假设表NEXT_DATE
中的COMPONENT_STATUS
为01-JAN-15
这是我目前写的SELECT
声明。
SELECT * FROM (SELECT c.* FROM
COMPONENTS c,
COMPONENTS_TRX ct
WHERE 1=1
AND ct.NAME = 'COMP_77C'
AND c.COMP_ID = ct.COMP_ID
AND c.COMP_ID IN (1,2)
AND trunc(c.DATE_FROM) < (SELECT trunc(ADD_MONTHS((NEXT_DATE),1)) FROM (SELECT NEXT_DATE FROM COMPONENT_STATUS
WHERE COMP_REF_CODE = (SELECT ID FROM REF_CODE WHERE CODE = 'COMP_ACTUAL')
ORDER BY NEXT_DATE DESC)
WHERE ROWNUM = 1
))
ORDER BY DATE_FROM DESC, COMP_TIMESTAMP DESC, COMP_TRX_ID DESC
;
它会从表格COMPONENT
返回所有内容,但正确排序。
如何使输出看起来像这样:
COMP_ID DATE_FROM TIMESTAMP COMP_TRX_ID
1 10-JAN-15 10-JAN-14 5
2 11-JAN-15 10-JAN-14 5
仅包含符合要求的第一行?
答案 0 :(得分:1)
您可以使用分组和聚合功能,具体取决于您的需求(如MAX
):
SELECT COMP_ID, MAX(DATE_FROM) AS DATE_FROM,
MAX(TIMESTAMP) AS TIMESTAMP, MAX(COMP_TRX_ID) AS COMP_TRX_ID
FROM (SELECT c.*
FROM
COMPONENTS c,
COMPONENTS_TRX ct
WHERE 1=1
AND ct.NAME = 'COMP_77C'
AND c.COMP_ID = ct.COMP_ID
AND c.COMP_ID IN (1,2)
AND trunc(c.DATE_FROM) < (SELECTtrunc(ADD_MONTHS((NEXT_DATE),1))
FROM (SELECT NEXT_DATE FROM COMPONENT_STATUS
WHERE COMP_REF_CODE = (SELECT ID
FROM REF_CODE
WHERE CODE = 'COMP_ACTUAL')
ORDER BY NEXT_DATE DESC)
WHERE ROWNUM = 1
))
GROUP BY COMP_ID
ORDER BY DATE_FROM DESC, COMP_TIMESTAMP DESC, COMP_TRX_ID DESC;
还要尝试避免使用逗号连接语法(已过时)并使用JOIN
:
FROM COMPONENTS c,
COMPONENTS_TRX ct
为:
FROM COMPONENTS c
JOIN COMPONENTS_TRX ct
ON c.COMP_ID = ct.COMP_ID