Re: [HACKERS] bug(?) if int8 as primary key - Mailing list pgsql-hackers

From Thomas G. Lockhart
Subject Re: [HACKERS] bug(?) if int8 as primary key
Date
Msg-id 3681048F.5BE33D7C@alumni.caltech.edu
Whole thread Raw
In response to bug(?) if int8 as primary key  (David Sauer <davids@iol.cz>)
List pgsql-hackers
> david=> create table t( i int8 primary key );
> NOTICE:  CREATE TABLE/PRIMARY KEY will create implicit index t_pkey 
>          for table t
> ERROR:  Can't find a default operator class for type 20.

int8 is a new data type, and I did not create the functions and catalog
entries required to implement a b-tree index yet. Since int8 relies on
compiler support for 8 byte integers, I wasn't sure how many platforms
would actually be able to compile the type. I think we know now that
most platforms do support the data type, so we should have more complete
support for it in the next release.

> If I restart pgsql, I can create table t (with int as primary key, not
> int8).

Yes, that's funny. It apparently does not quite clean up after itself.
The good news is that you only have to restart your session, not the
postmaster :)
                         - Tom


pgsql-hackers by date:

Previous
From: Oleg Bartunov
Date:
Subject: Re: [HACKERS] bug(?) if int8 as primary key
Next
From: Oleg Broytmann
Date:
Subject: Re: [HACKERS] Date/time on glibc2 linux