Thanks Bernt, that works!
However the only thing is that I've defined COLUMNS for the whole file 
using the #+COLUMNS keyword at the top of the file. I guess those 
properties are not being considered in the column view on the agenda 
buffer. It comes up with the default column properties.... I guess the 
only way is to redefine the standard column properties, so that the 
column properties that I want are displayed in the agenda view as well...

thanks,
Jose

Bernt Hansen wrote:
> Jose Robins <wulfhomme13-rook@yahoo.com> writes:
>
>   
>> Let's say I have a org tree with some tasks tagged with the keyword ":today:".
>> Now I do a sparse tree matching the today tag and this would limit the display
>> to only items with tag "today".
>>     
>
> Tag your tasks with :today: then do an agenda view tag match for today
>
> C-c a 1 m today RET
>
> Tag inheritance will control if subtasks are included or not.
>
>   
>> Now if I would like to do a column view on that sparse tree by doing C-c C-x
>> C-c. However this does not seem to limit the column display to only the
>> headlines with the tag "today".  Is it possible to achieve what I'm trying to
>> do? Is there any way to filter out (i.e hide) headlines in the column view
>> based on some criteria?
>>     
>
> Then C-c C-x C-c in the agenda view
>
> HTH,
> Bernt
>
>