Re: Yet again on indices...

Tom Lane wrote:<br /><blockquote cite="mid22262.1014823120@sss.pgh.pa.us" type="cite"><pre wrap="">Jean-Paul ARGUDO <a
class="moz-txt-link-rfc2396E"href="mailto:jean-paul.argudo@idealx.com"><jean-paul.argudo@idealx.com></a>
writes:<br/></pre><blockquote type="cite"><pre wrap="">EXPLAIN SELECT t12_bskid, t12_pnb, t12_lne, t12_tck<br />FROM
T12_20011231<br/>WHERE t12_bskid >= 1   <br />ORDER BY t12_bskid, t12_pnb, t12_tck, t12_lne;<br
/></pre></blockquote><prewrap=""><br /></pre><blockquote type="cite"><pre wrap="">Sort  (cost=3006.13..3006.13
rows=25693width=46)<br />  ->  Seq Scan on t12_20011231  (cost=0.00..1124.20 rows=25693 width=46)<br
/></pre></blockquote><prewrap=""></pre></blockquote> Try the following:<br /><br /> EXPLAIN ANALYZE SELECT t12_bskid,
t12_pnb,t12_lne, t12_tck FROM T12_20011231 WHERE t12_bskid >= 1  ORDER BY t12_bskid, t12_pnb, t12_tck, t12_lne;<br
/><prewrap=""></pre> and see what the actual results are.  Then turn the seq_scans off and do the same thing.<br
/><blockquotecite="mid22262.1014823120@sss.pgh.pa.us" type="cite"><pre wrap=""><br /></pre><blockquote type="cite"><pre
wrap="">=>Uh? Seq scan cheaper than index???  <br /></pre></blockquote><pre wrap=""><br />For that kind of query,
veryprobably.  How much of the table is<br />actually selected by "WHERE t12_bskid >= 1"?<br /><br />
regards,tom lane<br /><br />---------------------------(end of broadcast)---------------------------<br />TIP 2: you
canget off all lists at once with the unregister command<br />    (send "unregister YourEmailAddressHere" to <a
class="moz-txt-link-abbreviated"href="mailto:majordomo@postgresql.org">majordomo@postgresql.org</a>)<br
/></pre></blockquote><br/><br />