On 11/7/20 9:28 AM, Leandro Guimarães wrote:
> Hello,
> I have a process using pg_bulkload and sometimes i have duplicated
> keys in my csv file that pg_bulkload uses.
>
> My question is: pg_bulkload insert it in sequential order ?
>
> Example, if i have the following csv file:
>
> key_1;0.00
> key_1;100.00
>
> And use the ON_DUPLICATE_KEEP = NEW in .ctl file, it's guaranteed
> that the value 0.00 will be overwritten with 100.00? Or pg_bulkload
> can't guarantee this order?
Assuming they are in that order in the file and you are using DIRECT
mode I would say that would be the case. In PARALLEL mode, who knows?
In any case I would be dubious of any process that overwrites and
depends strictly on ordering to do the right thing. You are putting a
lot of confidence in the data in the CSV file being correctly ordered.
>
> Thanks!
> Leandro Guimarães
>
--
Adrian Klaver
adrian.klaver@aklaver.com
