On 8/10/19 7:47 PM, Souvik Bhattacherjee wrote:
> Hi Adrian,
>
> Thanks for the response.
>
> > Yes, but you will some code via client or function that batches the
> > inserts for you.
>
> Could you please elaborate a bit on how EXP 1 could be performed such
> that it uses bulk inserts?
I guess it comes down to what you define as bulk inserts. From your OP:
EXP 1: inserts with multiple txn:
insert into tab2 (attr1, attr2) (select attr1, attr2 from tab1 where
attr2 = 10);
insert into tab2 (attr1, attr2) (select attr1, attr2 from tab1 where
attr2 = 20);
If the selects are returning more then one row then you are already
doing bulk inserts. If they are returning single rows or you want to
batch them then you need some sort of code to do that. Something
like(pseudo Python like code):
attr2_vals= [(10, 20, 30, 40), (50, 60, 70, 80)]
for val_batch in attr2_vals:
BEGIN
for id in val_batch:
insert into tab2 (attr1, attr2) (select attr1, attr2
from tab1 where attr2 = id)
COMMIT
>
> Best,
> -SB
--
Adrian Klaver
adrian.klaver@aklaver.com
