Gullimeel (7/9/2012)
--------------------------------------------------------------------------------
I would suggest a derived column like diff in minutes since say Jan 01.01.1900 and put that columns in both the tables.Then use that column. for join along with the accountid.
Umm... Why? What relevance does the diff in minutes from 1900.01.01 have? To me, partitioning on date makes sense because the value is relevant and you can make a filegroup read only. Difference in minutes to an arbitrary date makes absolutely no sense to me.
Thanks,
This depends on what kind of query you would like to go against these tables. if he is going to write all the queries like date >=01/01/2012 and dat < 01/02/2012. Then just having a date column(just 3 bytes) makes much more sense.
But if he is quering say based on hours or minutes the adding that kind of column(int) makes much more sense. But it all depends on query...
GulliMeel
Finding top n Worst Performing queries[/url]
Improve the performance of Merge Join(special case)
How to Post Performance Problem -Gail Shaw[/url]