From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Gregory Stark <stark(at)enterprisedb(dot)com> |
Cc: | George Oakman <oakmang(at)hotmail(dot)com>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Indexing a Bit String column |
Date: | 2009-02-24 15:46:38 |
Message-ID: | 9271.1235490398@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Gregory Stark <stark(at)enterprisedb(dot)com> writes:
> Note that something like "WHERE myBitStringCol & B'101'" might be selecting
> too much of your table to make an index useful anyways. If each bit is set in
> half the table then you're talking about selecting 3/4 of the table in which
> case a full table scan would be more efficient than any index.
If the expectation is that the bitstring is mostly zeroes, I wonder
whether the OP wouldn't be better off using an integer-array
representation, ie instead of '00000101' store '{6,8}'. Then he could
use the existing GIST or GIN support for integer array operations.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | George Oakman | 2009-02-24 16:39:44 | Re: Indexing a Bit String column |
Previous Message | Gregory Stark | 2009-02-24 15:35:58 | Re: Indexing a Bit String column |