使用子查询更新列

Question

I am trying to update eid of the pay table with eid of payc table by joining payid .In the PayC table there are duplicate eid s' for matched payid s',Need to get only one EID for every matched PAYID

update PAY
SET eid = (select t.EID from (select a.EID,a.PAYID, count(a.EID) as cnt 
                                                       from PAYC a,PAYC b
                                                            where a.payid = b.payid
                                                            group by a.PAYID,a.eid
                                                            having count( a.eid) > 1) t
                                                            where  T.payid = PAYID)
WHERE EXISTS (select 'X' from (select a.EID,a.PAYID, count(a.EID) as cnt 
                                                       from PAYC a,PAYC b
                                                            where a.payid = b.payid
                                                            group by a.PAYID,a.eid
                                                            having count( a.eid) > 1) t
                                                            where  T.payid = PAYID)

Error:Subquery returned more than 1 value. This is not permitted when the subquery follows =, !=, <, <= , >, >= or when the subquery is used as an expression.

we can find duplicate EID by this query BUT NEED TO FETCH ONLY ONE EID

select a.EID,COUNT(a.PAYID), count(a.EID) as cnt 
 from PAYC a,PAYC b
 where a.payid = b.payid
 group by a.PAYID,a.eid
 having count( a.eid) > 1 AND 
 COUNT(a.PAYID) >1 

sample data

paycid payid  eid  amount year  quarter
1       101  1000   2000  2001   1
2       101  1000   3000  2001   2

Need to fetch

eid
1000

for the matched payid

Answer 1

select top 1 * from stuff

可能会。

Answer 2

Something like this might help?

WITH Duplicates AS (
    SELECT
        a.EID,
        a.PAYID,
        ROW_NUMBER() OVER (PARTITION BY a.PAYID ORDER BY a.EID) AS ROWID
    FROM
        PAYC a
        INNER JOIN PAYC b ON b.PAYID = a.PAYID
    GROUP BY
        a.EID,
        a.PAYID
    HAVING
        COUNT(a.EID) > 1)       
UPDATE
    p
SET
    EID = d.EID
FROM
    PAY p
    INNER JOIN Duplicates d ON d.PAYID = p.PAYID AND d.ROWID = 1;

Although I have to admit that I don't understand the logic of your self-join from PAYC to PAYC. Surely you don't need to do this to find cases where there are multiple EIDs per PAYID? In which case you could just remove the INNER JOIN PAYC b line entirely?

After thinking about this some more, could you not simply do this?

WITH Duplicates AS (
    SELECT
        PAYID,
        MAX(EID) AS EID
    FROM
        PAYC
    GROUP BY
        PAYID
    HAVING
        COUNT(DISTINCT EID) > 1)        
UPDATE
    p
SET
    EID = d.EID
FROM
    PAY p
    INNER JOIN Duplicates d ON d.PAYID = p.PAYID;

Then use MAX, MIN or whatever aggregation your require to pick the EID you want to use to update your PAY table (my example uses MAX). This is why your original query didn't work, you didn't tell it which EID to use, and it was returning multiple values.