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 |
|
NewMedia42
Starting Member
35 Posts |
 Posted - 2009-02-27 : 23:13:55
|
| I have two tasks that run and update the same table - the table gets about 6 million new records per day from the two processes. Currently I check before each insert to verify that an item matching it doesn't already exist - this is true probably > 80% of the time.My question is - would it be faster to just always insert (regardless if a match already exists, and just periodically run a task on the table to delete duplicates? This will obviously increase fragmentation, but is it worth exploring this? Looking through the profiler (code), it's spending 85% of it's time checking for pre-existing entries, so this would make a major speedup. |
|
|
sodeep
Master Smack Fu Yak Hacker
7174 Posts |
Posted - 2009-02-27 : 23:31:45
|
| All three(update/delete/insert) deals with external and internal fragmentation.You might wanna try with SSIS with Lookup transformation Tasks.There are plenty of Examples in SSIS topic. |
 |
|
|
darkdusky
Aged Yak Warrior
591 Posts |
|
|
|
|
|
Check before insert or periodically delete dupes?