On Mon, 14 Aug 2006, Mark Taylor wrote:

>  - Sending large catalogue row selections is currently shockingly
>    inefficient - O(N*M) for transmitting an M-row selection from 
>    an N-row catalogue - so could cause problems for large 
>    (>~1E4-row?) tables.  I can't see how to do these in a more 
>    sensible way, at least not without some fairly major reworking
>    of the AstroCat class or descendants.  Peter, maybe you could
>    take a quick look at PlasticSearch.tcl and see if you 
>    have any bright ideas (if it's not immediately obvious to you 
>    what the problem is let me know and I'll provide more commentary).

Apologies if you've got round to looking at this, I've realised I
can improve matters by using an associative array.  I'll fix it up.


Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
[log in to unmask] +44-117-928-8776