Curiosity with catalog table array attribute index dimensions

Erik Jones <erik@myemma.com> writes:

So, I noticed today that pg_index.indkey index values start at 0
instead of 1 as I'd expected. Are there other cases of this?
Shouldn't that be documented somewhere?

Well, the CREATE TYPE manual page has

For historical reasons (i.e., this is clearly wrong but it's far too
late to change it), subscripting of fixed-length array types starts from
zero, rather than from one as for variable-length arrays.

oidvector and int2vector aren't fixed-length anymore, but they maintain
their old behavior in this respect for backward compatibility.

regards, tom lane

On Apr 16, 2008, at 4:26 PM, Tom Lane wrote:

Erik Jones <erik@myemma.com> writes:

So, I noticed today that pg_index.indkey index values start at 0
instead of 1 as I'd expected. Are there other cases of this?
Shouldn't that be documented somewhere?

Well, the CREATE TYPE manual page has

For historical reasons (i.e., this is clearly wrong but it's far
too
late to change it), subscripting of fixed-length array types
starts from
zero, rather than from one as for variable-length arrays.

oidvector and int2vector aren't fixed-length anymore, but they
maintain
their old behavior in this respect for backward compatibility.

Ah, thanks, at least now I know what to look for instead of using
trial and error.

Erik Jones

DBA | Emma®
erik@myemma.com
800.595.4401 or 615.292.5888
615.292.0777 (fax)

Emma helps organizations everywhere communicate & market in style.
Visit us online at http://www.myemma.com