InstantForum.NET v2.9.0SQLServerCentralhttp://www.sqlservercentral.com/Forums/notifications@sqlservercentral.comFri, 24 May 2013 04:27:32 GMT20http://www.sqlservercentral.com/Forums/Topic1413740-360-1.aspxMaybe the clustered index that re-arranged the pages created disk fragmentation?Wed, 17 Apr 2013 10:49:56 GMTmmartin1http://www.sqlservercentral.com/Forums/Topic1413740-360-1.aspxThanks for your help.. BTW I put a covering index into place, but the job that inserts data into that table grew from 30min to 90min (give or take a few min). So I have removed the covering index, and the run time is back down. While the difference in run time of this query is only about 2min. Given that this query only runs once per day.. I am going to leave the covering index off.Fri, 01 Feb 2013 08:09:28 GMTdwilliscphttp://www.sqlservercentral.com/Forums/Topic1413740-360-1.aspxYou really should have both on a permanent table. Maybe as one and the same thing, maybe different.Thu, 31 Jan 2013 07:10:34 GMTGilaMonsterhttp://www.sqlservercentral.com/Forums/Topic1413740-360-1.aspxThanks... BTW it is an inner join. I have already sent a request off to the software company to see if we can create a PK. Currently there is not one. I will see about a clustered index, if we can not get a PK put into place.Thu, 31 Jan 2013 07:04:26 GMTdwilliscphttp://www.sqlservercentral.com/Forums/Topic1413740-360-1.aspxMaybe. Maybe not. Depends on the rows involved, the join type, other filters, complexity of the query, data types and a few other things. Add indexes, see if they help.Temp tables are not necessarily memory resident, but anything that they query processor operates on will be in memory, temp or permanent, the query processor doesn't know what a disk is.Ignoring joins, the permanent table needs a clustered index.Wed, 30 Jan 2013 12:35:38 GMTGilaMonsterhttp://www.sqlservercentral.com/Forums/Topic1413740-360-1.aspxI am fuzzy on this.. so looking to see if my thoughts are correct... When looking at the execution plan, it does not ask for an index, but...If you have two moderate (500K to 1M) tables, and both are a heap, wouldn't it help the inner join to have index's on what is being joined? If it matters, one is a temp table (#temp) so it would already exist in memory.. one would hope.Wed, 30 Jan 2013 12:22:54 GMTdwilliscp