Indeed, but .....
this means that the standard VOTable readers then cannot assume that there is the mandatory tag. So they have to be able to parse without it, thus why bother at all?
To have the row count in the end is still helpful, because we can compare, it serves as a checksum.
Cheers, Alex
-----Original Message-----
From: owner-votable-at-eso.org [mailto:owner-votable-at-eso.org]On Behalf Of
Alasdair Allan
Sent: Thursday, October 23, 2003 8:50 AM
To: Alex Szalay
Cc: Patricio F. Ortiz; Alasdair Allan; Randall Thompson;
votable-at-ivoa.net
Subject: RE: VOTable Questions
> A small add-on info: with SDSS, people are routinely requesting output
> tables containing over 10 million rows. This is the typical size that we
> should be worried about, not tables with 100-1000 rows.
Yes, agreed...
> I like Guy's suggestion, that would avoid any conflicts with streaming.
While I was wavering, I've now been convinced that Guy's suggestion doesn't help because to get to the meta-data you'd have had to parse the entire TABLEDATA element anyway.
So my current vote is an optional attribute to TABLEDATA, after all the number of rows in a TABLEDATA should be associated as closely as possible with the TABLEDATA itself? Putting it in an element "elsewhere" looses that association.
Of course this doesn't help you in the streaming case, but with a streaming table you've committed yourself to parse the entire thing anyway so you wouldn't be paying any attention to the tag anyway...
Al. Received on 2003-10-23Z18:13:07