Rick,
I am glad I could help, but I am not quite sure you understand the
purpose/use of the trigger to
partition the table.
The trigger merely decides which child should get the data, the "query" is
not passed, only the data,
To be more specific, if your appropriate child table were child_1000, t
Melvin, thanks. Syntax was indeed the problem there. Any idea how to pass the
original query to the child table? My insert has 113 parameters passed in,
but based on the constraint violations I am seeing, it seems they are not
making. I suspect that: VALUE (NEW.*) does not pass in the original quer
Hi Jim,
I'm not going the partitioning route because I want to, I'm just out of
options at this point. As this table gets bigger, performance just gets
worse over time. I wanted to try partitioning to see if it helps.
Thanks for the tip, looking at the function again what you suggest makes
perfec
On 4/29/15 10:05 AM, akshunj wrote:
IF ('invoice' <= 'I-1') THEN INSERT INTO myschema.mywork VALUES (NEW.*)
ELSE IF ('invoice' >= 'I-10001' AND <= 'I-2'
That's going to fall apart with invoice I-10.
If you're going to go this route, depend on how IF ELSIF operates and
don't try to
I think the problem is you need to specify NEW.invoice in all comparisons
(don't quote the 'column')and always, Always, ALWAYS end each statement
with a semicolon.
IE:
IF (NEW.invoice <= 'I-1') THEN INSERT INTO myschema.mywork VALUES
(NEW.*)
ELSE IF (NEW.invoice >= 'I-10001' AND NEW.invoice <=
Hi,
I am trying to setup partitions and as a test, I was able to follow the
example in the Postgres docs using the date as a condition.
Now I am trying to partition on a column with the data type character
varying. I want to partition based on an invoice ID consisting on letters
and numbers like
