On Fri, 30 Nov 2001 10:34:11 +0000, Tim Bunce wrote:
>On Thu, Nov 29, 2001 at 03:08:32PM -0500, Charles Day wrote:
>> Might be slightly OT, sorry. I've been researching this for some time now.
>> Exactly like search engines. Past ideas were to create a temporary db or
>> use cursors. Anybody have ideas?
>
>http://search.cpan.org/doc/TIMB/DBI_Talk5_2001/sld059.htm
You've got some pretty good ideas up there... My favourite is "Select
and cache only the row keys, fetch full rows as needed".
But you might not need the cache, then. When combined with the first
idea, "re-execute query and discard unneeded", this can become:
Re-execute full query only fetching the row id's.
Count and discard (this can be done in a perl array),
and then fetch the full rows only for those rows you want.
--
Bart.
