我有一个需要分组的数据列表,但我们只想对计数大于3的数据进行分组。
AA
AA
BB
CCC
CCC
CCC
返回
AA 1
AA 1
BB 1
CCC 3
感谢您的帮助
答案 0 :(得分:3)
select data, case when total < 3 then 1 else total end total
from
(
select data, Count(Data) Total
from tbl
group by data
) g
join (select 1 union all select 2) a(b)
on a.b <= case when total < 3 then Total else 1 end
order by data
在大多数情况下,这应该比LittleBobbyTables的答案更快。
答案 1 :(得分:2)
在我的脑海中,您可以使用计数大于2的所有内容,然后使用UNION ALL
获取不在第一个查询中的任何记录:
SELECT 'AA' AS Data
INTO #Temp
UNION ALL SELECT 'AA'
UNION ALL SELECT 'BB'
UNION ALL SELECT 'CCC'
UNION ALL SELECT 'CCC'
UNION ALL SELECT 'CCC'
SELECT Data, COUNT(Data) AS MyCount
FROM #Temp
GROUP BY Data
HAVING COUNT(Data) > 2
UNION ALL
SELECT Data, 1
FROM #Temp
WHERE Data NOT IN (
SELECT Data
FROM #Temp
GROUP BY Data
HAVING COUNT(Data) > 2
)
ORDER BY Data
DROP TABLE #Temp
答案 2 :(得分:0)
使用窗口函数:
select col, count(*) as cnt
from (select col, count(*) over (partition by col) as colcnt,
row_number() over (order by (select NULL)) as seqnum
from t
) t
group by col, (case when colcnt < 3 then seqnum else NULL end)
计算列的总计数和每行的唯一标识符。然后group by子句测试条件。如果小于3,则它使用标识符来获取每一行。如果更大,则在这种情况下使用常量值(NULL)。