From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | khalil(dot)alijani(at)gmail(dot)com |
Cc: | pgsql-bugs(at)postgresql(dot)org |
Subject: | Re: BUG #6463: tables can have at most 1600 columns |
Date: | 2012-02-18 16:30:22 |
Message-ID: | 27713.1329582622@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-bugs |
> I drop a column but Postgresql still shows old (deleted) columns...
> My Java program does this:
> drop a boolean column and then add same column (with same name and
> specification), in dozens times, because updating (this) column is very slow
> !!!
> AND finally it says "tables can have at most 1600 columns"
Sorry, this is not a bug, or at least we're not going to fix it.
The fact that DROP COLUMN doesn't physically remove the column is an
intentional design decision. If we didn't do it that way, dropping
a column would require rewriting the entire table (and all its indexes)
and so would be horribly slow.
Perhaps if you discussed why you think that repeatedly dropping and
re-adding a column is a useful thing to do, we could help you find
another way. The bugs list is not the place for that discussion,
however.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | crees | 2012-02-18 17:35:05 | BUG #6465: postgresql-pltcl refuses to compile with 9.1, and FreeBSD does support threaded python |
Previous Message | Tom Lane | 2012-02-18 16:21:06 | Re: BUG #6462: rpmbuild fails during regression tests on SMP ARM machines |