[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Support #TVU-725269]: Is there any output similar to what pqbinstats created?

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [Support #TVU-725269]: Is there any output similar to what pqbinstats created?
  • Date: Tue, 20 May 2008 09:08:55 -0600
Gregg,

> Thanks for the details.  I'm curious now what entry I need to put in
> pqact.conf to file the statistics being generated and inserted into the
> LDM queue?  I searched the Unidata website and googled for a while and
> it seems most LDM users just send the stats back to Unidata
> via the -h option in rtstats.  I did find:
> 
> http://www.unidata.ucar.edu/support/help/MailArchives/ldm/msg03326.html
> 
> Where Steve Chiswell indicated to Carl Sinclair the following:
> 
> >EXP     ^stats/(.*)/(.*)/(.*)
> >        FILE    data/shared/rtstats/\2/%Y%m%d/\1/\3
> 
> The pattern for LDM6 will be
> 
> ^rtstats-LDMVERSION/LDMHOST/FEEDTYPE/PRODUCT_ORIGIN
> 
> The manpage also references LDMVERSION, LDMHOST, FEEDTYPE,
> PRODUCT_ORIGIN as well.
> 
> Can you please point me to documentation on how to effectively store the
> data generated by rtstats (e.g. pqact.conf configuration entry)?

The effectiveness of the manner in which the data is stored depends
on how the data will be accessed and used.  The rtstats(1) plots on
our website are generated by GEMPAK.  Because GEMPAK can read netCDF
files, we store the data in that form.

It's unlikely that our solution would be appropriate for you.

If I were to start over from scratch, I might use gnuplot(1) rather
than GEMPAK to plot the data.  In which case, I would probably store
the data in columns in textual form, which is easy because the
data-products are, themselves, column-oriented text.

The structure of the product-identifier seems appropriate for
a directory structure.  Because the product-identifier doesn't
encode time, I would probably use pqact(1)'s ability to use the
data-product creation time in generating the name of the file to
which to append the data-product.  All products for a given hour
could be appended and the files could be purged via scour(1) and
a crontab(1) entry.

> Thanks,
> Gregg

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: TVU-725269
Department: Support LDM
Priority: Normal
Status: Closed