DSXchange

Hi folks, I am trying to use the Bulk Load option in the Oracle Connector stage.

I have found the following options are required to get a successful load:

1) Use before SQL statement to remove pk and and drop all indexes (including pk index).

2) Use after SQL statement to create pk and all indexes.

3) Set Index maintenance option to "Skip All"

4) Disable logging.

5) Perform operations before/after bulk load? No.

I am wondering if there is anything wrong with this. It seems to be working fine, it reduces my load time from 32 minutes to 14 minutes, although I haven't checked the db to see if it is actually using Direct Load.


Thanks,
Doug

It's OK provided that you guarantee - totally - that the data you are loading will not violate any constraints (typically uniqueness constraints in the case of PK index). You may also want to re-calculate table statistics after performing all these tasks.

ray.wurlod wrote:It's OK provided that you guarantee - totally - that the data you are loading will not violate any constraints (typically uniqueness constraints in the case of PK index). You may also want to re-calculate table statistics after performing all these tasks.

Thanks Ray.

Doug