Pg_bulkload loading about pg_bulkload HOT 6 CLOSED

Hi,

-->Pg_bulkload is more like Sql loader in Oracle right? What if the table has indexes. Will it be rebuild after loading data as like oracle? Not sure, how this process works.

Yes, the indexes are rebuilt after loading data, not to updating indexes in insertition.

-->In pg_bulkloads, records will be inserted not updated. right?
Came across, few records got replaced. Is it because of ON_DUPLICATE_KEEP = NEW. Will that be be updated, or record will be removed and re-inserted. Kindly clarify that.

Yes, load records are inserted, not updated.
In default configuration,ON_DUPLICATE_KEEP does not work because DUPLICATE_ERRORS is non zero.
Could you tell us the situation more detail?

from pg_bulkload.

Thanks.

This is my case, where records got replaced and i have a settings as

DUPLICATE_ERRORS = 0
ON_DUPLICATE_KEEP = NEW

This is normal.right? If any record was already present, that particular record will be replaced by a record from source.

I have one more question. Will pg_bulkload disable the trigger before loading? I don't want to disable the trigger.

from pg_bulkload.

This is normal.right? If any record was already present, that particular record will be replaced by a record from source.

Yes, that is the default configuration.
I think that pg_bulkload will abort if there is duplicated records in this configuration like below.

$ pg_bulkload sample_csv.ctl 
NOTICE: BULK LOAD START
ERROR: query failed: ERROR:  could not create unique index "foo_pkey"
DETAIL:  Key (a)=(1) is duplicated.
DETAIL: query was: SELECT * FROM pg_bulkload($1)

I have one more question. Will pg_bulkload disable the trigger before loading? I don't want to disable the trigger.

Yes, pg_bulkload does not call triggers.

from pg_bulkload.

Thanks.

We created partitioned table based on month or day.

Based on the process date in source file, trigger will look whether the partitioned was already present, if not, will create a partition and load the data.That's essential from our side.

Any other workaround for this using pg_bulkload.

from pg_bulkload.

Because pg_bulkload does not use triggers, you should prepare data files for each partition tables and load each files into each partition tables by pg_bulkload.

from pg_bulkload.

Thanks

from pg_bulkload.