Re: Is indexing broken for bigint columns?

> -----Original Message-----
> From: Mike Mascari [mailto:mascarm(at)mascari(dot)com]
> Sent: Tuesday, February 24, 2004 3:27 PM
> To: Dann Corbit
> Cc: PostgreSQL-development
> Subject: Re: [HACKERS] Is indexing broken for bigint columns?
>
>
> Dann Corbit wrote:
> > http://www.phpbuilder.com/columns/smith20010821.php3?page=3
> >
>
> bigint indexes work fine. The queries probably referenced 32-bit
> integer constants that were neither quoted nor CAST. I always start
> bigint sequences at 5 billion. This ensures that client applications
> aren't assuming 32-bit quantities that will break once ~4.2 billion
> is reached and I get index scans without quoting or casting free.
> But IIRC there's a change in the development tree to jettison the
> requirement for quoting/casting...

I think it would be awfully nice for "conversions that make sense" to
happen implicitly.
Including for function calls.
Including comparison operators.
Etc.

I have had to write a ton of work-around stuff and I still keep finding
holes in it.