Re: max # of columns in the table (again!)

From: Tanel Poder <tanel_at_remove.this.peldik.com>
Date: Sat, 15 Feb 2003 11:18:27 +0100
Message-ID: <3e4e0652$1_1@news.estpak.ee>

Sybrand,

since in the row format there's only one byte for column count, chaining within the same block is used to have another 255 columns, so if we'd have 600 column row in a block, it'd be chained twice to the same block, if block space permits.

Tanel.

"Sybrand Bakker" <gooiditweg_at_nospam.demon.nl> wrote in message news:qq0s4v8m29tp3rhmmll5q3tet93codfqhl_at_4ax.com...
> On 14 Feb 2003 23:15:58 -0800, ydias_at_hotmail.com (dias) wrote:
>
> >Oracle (8i) has an internal limit of 255. That means that if any
> >columns beyond 255 are not null, that row will always be chained.
>
> Sounds like nonsense to me. Whether a row is chained is determined
> *ONLY* by the number of bytes available in a block, NOT by the number
> of columns and NOT by a FIXED limit.
> In fact, it is quite possible, especially when you choose a small
> block size, to have rows with long varchar2s chained.
>
>
> Sybrand Bakker, Senior Oracle DBA
>
> To reply remove -verwijderdit from my e-mail address
Received on Sat Feb 15 2003 - 04:18:27 CST