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 |
|
niranjankumark
Posting Yak Master
164 Posts |
 Posted - 2007-12-01 : 06:03:56
|
| Hi,I have four tables.Every week end i will truncate the old data and populate the new data.in this table i will do some validation based on 3 or 4 columns.It takes more time, because there is no index.If i put nonclustered index will it improve performance????because i will delete the entire data and insert new on next week.in this deletion and insertion scenario, creating non cluster will improve performance???? |
|
|
TG
Master Smack Fu Yak Hacker
6065 Posts |
Posted - 2007-12-01 : 08:14:41
|
>>If i put nonclustered index will it improve performance????>>It takes more time, because there is no indexsounds like you answered your own question  It is impossible to answer your question without knowing any details about your table structure and the specific validation you are performing. If you want answers, post the ddl and validation code.Be One with the OptimizerTG |
 |
|
|
niranjankumark
Posting Yak Master
164 Posts |
Posted - 2007-12-04 : 01:44:17
|
| For a normal transaction table index will improve performance.But i am in situtation like...1020 columns in my table.Every week i have to populate arround 3 lakhs record on this table.before my population old records(3 lak) will get truncated.index shud not create this kind of large truncation table????because based on this table(which has no index) i am doing calculation.so the performance is very low without index. |
|
|
|
rmiao
Master Smack Fu Yak Hacker
7266 Posts |
Posted - 2007-12-04 : 23:03:57
|
| You can drop indexes before loading data then recreate them afterwards. |
|
|
|
TG
Master Smack Fu Yak Hacker
6065 Posts |
Posted - 2007-12-05 : 08:09:48
|
| An index (or lack of one) may have nothing to do with your slow performance. You say your table has 1020 colulmns?? That doesn't sound like a very good design. It is way too many columns. I don't know how many rows "3 lakhs record" means but as I said before, we need to see specifically what validations you are performing to be able to advise you on improving performance.What is your table structure?What validation are you performing?how many rows are in the table?how long is the validation currently taking?Be One with the OptimizerTG |
|
|
|
SwePeso
Patron Saint of Lost Yaks
30421 Posts |
Posted - 2007-12-05 : 08:16:05
|
300 000 records.
E 12°55'05.25"N 56°04'39.16" |
|
|
|
|
|
|
Non cluster