Re: heaptuple over the network

From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: Mon Nsi <wong(dot)fei(at)yahoo(dot)com>
Cc: Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: heaptuple over the network
Date: 2006-03-04 21:39:38
Message-ID: 20060304213938.GH13230@surnet.cl
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Mon Nsi wrote:

Hi,

> My question, and I am not a DB expert, I just deal
> with raw data. What other information do I need to
> include from struct HeapTupleData in order for that
> data to make sense on the other end (to rebuild a
> HeapTuple from it).
>
> I understand that:
>
> data_address = offset(HeapTupleHeaderData,
> tuple->t_data->t_hoff);
> data_len = tuple->t_len - tuple->t_data->t_hoff;
>
> tuple->t_data->t_ctid and tuple->t_data->t_natts are
> needed for this. Is that it?
>
> What do I need to watch for with TOAST'ed tuple?

You need to detoast any attribute that is stored externally.
A compressed attribute will work fine, but an externally stored
attribute just won't work because it references the toast table by OID,
which will not be available on the remote system.

You also need to send the null bitmap. OTOH I don't think you need to
CTID.

Not sure if anything else.

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Alvaro Herrera 2006-03-04 21:41:49 Re: Not so happy with psql's new multiline behavior
Previous Message Michael Paesold 2006-03-04 20:54:41 Re: Not so happy with psql's new multiline behavior