Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
Hariarul
Posting Yak Master
160 Posts |
Posted - 2006-11-30 : 00:57:00
|
| Hi all, My question is regarding the SQL server 2005 table partitions.The scenario:=============I have 2 huge tables A & B with table A having 1 million records and table B 2 million records. When I run a procedure which has few computations involving tables A & B, the time taken for the stored procedure to return the result set is 55 seconds.Now, I partition table A based on a column which is used in the join condition within the stored procedure and then try the same query.Still, it takes the same 55 seconds and I could not find any time differences even after going for partitioning and placing the partitions in different file groups.Clarifications needed:======================1. Is this approach correct?2.Have anyone done partitions before and noticed significant performance improvements?3.Should I also consider any other factors when doing partitions?Thanks & Regards,HHA |
|
|
Kristen
Test
22859 Posts |
Posted - 2006-11-30 : 01:44:47
|
| I don't think that 3 million rows is big.If you partition your table and you query only covers the range in the "1 million" part, then the query can be quicker. However, if the original table was indexed such that the query used the index it would be the same speed in both cases.A query on 3 million rows that takes 55 seconds suggests that the query needs optimising!Suggest you post the query here - and DDL of the tables etc. - so people can make suggestions.Kristen |
 |
|
|
|
|
|