DSXchange

Hi,

wht are the possible ways to improve performance when doing Oracle table updates.

i thought of Indexes, splitting up inserts and updates..

let me know any ways to improve the performace, my update job is taking longer time to update the records

any help is appreciated.

thanks

kool78 wrote:Hi,

wht are the possible ways to improve performance when doing Oracle table updates.

i thought of Indexes, splitting up inserts and updates..

let me know any ways to improve the performace, my update job is taking longer time to update the records

any help is appreciated.

thanks

Delete and then bulk load can be a option. Again if Oracle you can truncate a partition and load a new one if the load is on a monthly basis. Actually there are a lot of ways. Please provide more info on the background.

Thanks
Avi

avi21st wrote:

kool78 wrote:Hi,

wht are the possible ways to improve performance when doing Oracle table updates.

i thought of Indexes, splitting up inserts and updates..

let me know any ways to improve the performace, my update job is taking longer time to update the records

any help is appreciated.

thanks

Delete and then bulk load can be a option. Again if Oracle you can truncate a partition and load a new one if the load is on a monthly basis. Actually there are a lot of ways. Please provide more info on the background.

Thanks
Avi

we have a dimension which has a load_date column, so almost all the rows of the dimension gets updated daily, its a diaily run.

kool78 wrote:

avi21st wrote:

kool78 wrote:Hi,

wht are the possible ways to improve performance when doing Oracle table updates.

i thought of Indexes, splitting up inserts and updates..

let me know any ways to improve the performace, my update job is taking longer time to update the records

any help is appreciated.

thanks

Delete and then bulk load can be a option. Again if Oracle you can truncate a partition and load a new one if the load is on a monthly basis. Actually there are a lot of ways. Please provide more info on the background.

Thanks
Avi

we have a dimension which has a load_date column, so almost all the rows of the dimension gets updated daily, its a diaily run.

What is the database?

avi21st wrote:

kool78 wrote:

avi21st wrote: Delete and then bulk load can be a option. Again if Oracle you can truncate a partition and load a new one if the load is on a monthly basis. Actually there are a lot of ways. Please provide more info on the background.

Thanks
Avi

we have a dimension which has a load_date column, so almost all the rows of the dimension gets updated daily, its a diaily run.

What is the database?

Oracle 10g

kool78 wrote:

avi21st wrote:

kool78 wrote:
we have a dimension which has a load_date column, so almost all the rows of the dimension gets updated daily, its a diaily run.

What is the database?

Oracle 10g

Delete the updated records...then bulk load both insert+updates to the database works faster than updates. Else put them in Sequential file and do a Oracle import (SQL wrapped in shell script). What is the data volume and the time taken

Run Oracle on bigger hardware - mainframe perhaps.