Carlos Henrique Reimer writes:
> Yes, once correct schema was included in the search_path, VACUUM and ANALYZE
> run fine again.
You'd be better advised to fix the function so it works regardless of
caller's search_path. As-is, it's a loaded gun pointed at your foot.
rega
Hi,
Yes, once correct schema was included in the search_path, VACUUM and ANALYZE
run fine again.
Thank you!
On Fri, Sep 10, 2010 at 11:38 AM, Tom Lane wrote:
> Carlos Henrique Reimer writes:
> > Yes, you're right! I found out a functional index using this function and
> > ANALYZE also cancels
Carlos Henrique Reimer writes:
> Yes, you're right! I found out a functional index using this function and
> ANALYZE also cancels.
> Is there a way to code this function in a way VACUUM/ANALYZE does not
> cancel?
I think your problem is probably a search_path issue, ie vacuumdb is not
running wi
Hi,
Yes, you're right! I found out a functional index using this function and
ANALYZE also cancels.
Is there a way to code this function in a way VACUUM/ANALYZE does not
cancel?
Thank you!
brasil=# analyze "BRASIL".tt_tit;
ERROR: invalid type name "TT_TIT.SEQCAN%TYPE"
CONTEXT: compile of PL/p
Carlos Henrique Reimer writes:
> We are facing the following problem in a PG 8.2 server when trying to vacuum
> one of our databases:
> vacuumdb: vacuuming of database "reimer" failed: ERROR: invalid type name
> "TT_TIT.SEQCAN%TYPE"
> [ which seems to be coming from out-of-date code in a function
Hi,
We are facing the following problem in a PG 8.2 server when trying to vacuum
one of our databases:
vacuumdb: vacuuming database "reimer"
INFO: vacuuming "pg_catalog.pg_database"
INFO: "pg_database": found 0 removable, 6 nonremovable row versions in 1
pages
INFO: index "pg_database_datname_
