1

In my user table, I have a userid column, and membergroupids column. Here is the query I have so far.

UPDATE `user`
SET `membergroupids`=`,100`
WHERE `userid` IN
(
  312 , 331 , 538 , 528 , 77 , 80 , 293 , 343 , 609 , 607 , 547 , 596 , 216 ,
  780 , 1 , 884 , 430 , 500 , 881 , 952 , 957 , 979 , 352 , 994 , 955 , 776 ,
  744 , 904 , 1038 , 76 , 1083 , 61 , 202 , 1001 , 1173 , 1176 , 418 , 770 ,
  1192 , 191 , 1211 , 1206 , 1247 , 402 , 423 , 584 , 1233.1282 , 167 , 1329,
  1330 , 376 , 1227 , 146 , 1113 , 1327 , 1352 , 254 , 1065 , 1406 , 1225 ,
  1230 , 1405 , 1443 , 812 , 1295 , 1510 , 1571 , 1565 , 1468 , 1349 , 1645 ,
  501 , 1312 , 1718 , 13 , 200 , 1177 , 1785 , 107 , 540 , 1591 , 1508 , 59 ,
  1867 , 222 , 1520 , 782 , 26 , 332 , 474 , 1837 , 1828 , 1431 , 1951 , 1953 ,
  1950 , 847 , 1477 , 1726 , 1669 , 1185 , 1955 , 1872 , 1926 , 1392 , 853 ,
  1892 , 2020 , 1511 , 2027 , 2197 , 2184 , 2134 , 2239 , 119 , 122
)

So basically, what I am trying to do is.. I am trying to make it so the query adds ",100" to the column membergroupids if their userid is one of the ones listed in ().
However, here is the thing. All of these users already have values listed in membergroupids.
For example, the user 312 has these values already in the column membergroupids:
35,49,58,61,71,38,59
So, this query needs to do this to his field:
35,49,58,61,71,38,59,100

..and for all of the other userids listed.

Thanks if anyone can help me. I'm sure this is on google but I didn't know how to word it.

4

2 回答 2

2

Since you're in the unfortunate position of not being able to change the schema, you can add the value by concatenating it onto the existing values via CONCAT()

UPDATE `user` SET `membergroupids` = CONCAT(`membergroupids`, ',100') WHERE...

That will force ,100 onto every user, even if they currently have an empty membergroupids. To account for the empty ones, you can use a CASE condition:

UPDATE `users`
SET `membergroupids` =  
  /* Concatenate ,100 onto the non-empty ones */
  CASE WHEN `membergroupids` IS NOT NULL AND `membergoupids` <> '' 
    THEN CONCAT(`membergroupds`, ',100')
  /* And set the empty ones to just 100 (no comma) */
  ELSE '100'
  END
WHERE ...

In an ideal world, the user <--> group memberships would be handled with a proper many-to-many relationship, whereby a third table is used to assign users to groups, having two columns: userid, groupid.

于 2013-03-30T14:27:56.840 回答
1 
UPDATE `user` SET `membergroupids`=CONCAT(`membergroupids`, ',100') 
WHERE `userid` IN ( ... )

Of course this assumes you have at least one membergroupid in the list already. Appending ',100' if the list is empty results in a list with a leading comma, which is not usually desirable. Here's a quick solution:

UPDATE `user` SET `membergroupids`=CONCAT_WS(COALESCE(`membergroupids`, ''), '100') 
WHERE `userid` IN ( ... )

Also what happens if the list of membergroupids grows too long, and can't fit in the length limit for the varchar column? "I'm sorry, you can't join that group because we can't store your list of group memberships. Perhaps you'd like to join some groups with shorter values for their group ids, then you could join more groups."

Storing lists of values in a comma-separated string is a bad design for a relational database. See a description of pitfalls in my answer to Is storing a delimited list in a database column really that bad?

It would be easier to add a membergroup for a user if you stored the memberships one per row in an intersection table. Then you don't need to know the current list of values.

INSERT INTO user_membergroups (userid, groupid) 
SELECT `userid`, 100 FROM `user`
WHERE `userid` IN ( ... )
于 2013-03-30T14:30:15.863 回答