On Tue, Mar 09, 2004 at 01:47:40PM -0800, Bill Moseley wrote:
> On Tue, Mar 09, 2004 at 01:33:58PM -0800, Peter Ensch wrote:
> > Is there a way to query swish so that I can build the section 
> > list on the fly? IE. have it tell me what sections were captured
> > by ExtractPath and stored in the 'section' metaname.
> 
> Not in a very nice way.  What you are asking it to dump the index file
> and see what words are associated with the metaname you are using for
> ExtractPath.
> 
> So, assuming you are using "site" as the metaname:
> 
>   $ swish-e -T index_metanames | grep site
> 
> Then use
> 
>   $ swish-e -T index_words_meta
> 
> and grep out the id (the second column) that matches the "site" meta id
> number.
> 

Bill,

Thanks for the reply.
This will work; I can dump the index and parse out the words 
associated w/ a metaname at the time I index and use the output 
to build my form widgets.

I am concerned w/ the output however. The word names appear
truncated. I guess they are printf'ed for formatting purposes,
but this does limit their usefulness. Is there a way to 
prevent this w/out hacking on the source code?

An example of what I'm talking about is below.

Thanks,
Peter

---
ExtractPath section regex !^/path/to/my/site/([^/]+)/.*$!$1!

%> ls -1 /path/to/my/site
cgidata
design_manual
design_methodology
efuse
getstr
io_functions
legal
mls
pdf
pyramid
relnotes
test

%> swish-e -T index_words_meta | egrep '13$'
cgidata 13
design  1       10      13
efus    1       10      13
getstr  13
io      1       10      13
legal   1       13
manual  1       10      13
methodologi     1       10      13
ml      1       10      13
other   1       10      13
pyramid 1       10      13
relnot  13
test    1       10      13

-- 

^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^
Peter Ensch,
pensch@ti.com           A-1140   (214) 480 2333
^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^