Thread: AW: [HACKERS] now 6.4

AW: [HACKERS] now 6.4

From
Andreas Zeugswetter
Date:
> Yep, I think this is do'able, UNLESS Vadim decides he needs to change
> the structure of the data/index files.  At that point, we are lost.

> In the past, we have made such changes, and they were very much needed.
> Not sure about the 6.4 release, but no such changes have been made yet.

I thought Vadim was going to change the oid in btree index files to ctid,
in my opinion a very useful change. (Or was he intending to add it ?)
Then a btree index rebuild would be necessary.

Andreas



Re: AW: [HACKERS] now 6.4

From
Vadim Mikheev
Date:
Andreas Zeugswetter wrote:
>
> > Yep, I think this is do'able, UNLESS Vadim decides he needs to change
> > the structure of the data/index files.  At that point, we are lost.
>
> > In the past, we have made such changes, and they were very much needed.
> > Not sure about the 6.4 release, but no such changes have been made yet.
>
> I thought Vadim was going to change the oid in btree index files to ctid,
> in my opinion a very useful change. (Or was he intending to add it ?)
> Then a btree index rebuild would be necessary.

OID was removed from btree tuples ~ year ago.
Now I want to use heap tuple ID (referenced by index tuple)
as (last) part of index key and get rid of BT_CHAIN flag:
all keys will be UNIQUE and there will be no problems with
handling duplicate keys any more (idea (C) Oracle -:)

But this means that heap tuple id will be added to index
tuples on internal pages, not on the leaf ones...

Vadim