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

[THREDDS #KKM-141749]: offsetsGreaterEqual



Hi David,

I was able to reproduce what you are seeing locally, and I think it is a bug.  
I need to check with the developer who worked on this functionality to make 
sure I haven't made an error in setting up the fmrc (or made an error in 
understanding what the offsetsGreaterEqualshould do), but this issue will 
probably end up in our jira tracking system as a bug.  I'll send the 
information on the tracking ticket to you when I enter it in.

Thank you for identifying this phenomonon, I'm sure it was a subtle but 
frustrating observation on your end.

Have a good Memorial Day Weekend!

Cheers,
  Lansing

> Hi,
> 
> I have added a near (no ConstantForecasts) complete featured FMRC at:
> 
> http://tds.marine.rutgers.edu:8080/thredds/catalog/roms/espresso/2013_da/fmrc_t1/his/catalog.html
> 
> That dataset is with the 26 hour offset for Best. You should be able to
> reproduce the behavior by retrieving the *_20130505_* and *_20130506_*
> files and mimicking my configuration at the bottom of this email.
> HTTPServer access is enabled for the individual files.
> 
> When we tested the dataset we compared the variable zeta in the first
> record from the "Best" dataset to zeta in the first record of file
> espresso_his_20130505_0000_0002.nc and the second record of file
> espresso_his_20130506_0000_0001.nc.
> 
> We expected the first record of file espresso_his_20130505_0000_0002.nc
> to match but instead the second record from
> espresso_his_20130506_0000_0001.nc matched.
> 
> Please let me know if you need any other information.
> 
> Thanks,
> Dave
> 
> 
> 
> 
> On 5/22/13 2:16 PM, Unidata THREDDS Support wrote:
> > David,
> >
> > Can bundle these files up and send them to us, or point us to where we can 
> > grab them?  We'd like to reproduce the behavior locally for testing and 
> > debugging.
> >
> > -Lansing
> >
> >>>>>>>>>
> >
> > Hello,
> >
> > I am running the latest stable versions (as of May 20, 2013) of Java 
> > (1.7.0_21), Tomcat(7.0.40) and TDS(Version 4.3.16 - 20130319.1353).
> >
> > I have 7 day hindcast/forecasts that run once daily with output every 2 
> > hours that I am trying to put into an fmrc. My files are named using the 
> > the start date/time of the hindcast. I want to exclude the first 24 hours 
> > of each run from "Best" datasets. I thought using offsetsGreaterEqual="26" 
> > would do the trick but it only excludes the first 24 hours from run 0 and 
> > not from subsequent runs. Meaning, if I look at hour 26 of the "Best" 
> > dataset (the first record in that dataset) I'm getting data from the 
> > beginning (first 24 hours) of run 1 instead of hour 26 from run0.
> >
> > Is this the intended behavior of offsetsGreaterEqual? If so I think the 
> > documentation at:
> >
> > http://www.unidata.ucar.edu/projects/THREDDS/tech/tds4.3/reference/collections/FmrcCollection.html
> >
> > should be clarified. Otherwise I'd like to report a bug.
> >
> > Thanks,
> > David Robertson
> >
> >
> >
> > Catalog entry:
> >
> >     <featureCollection name="espresso_2013_da_history_fmrc"
> >                        featureType="FMRC"
> >                        harvest="true"
> >                        path="roms/espresso/2013_da/fmrc/his">
> >        <metadata inherited="true">
> >           <serviceName>both</serviceName>
> >        </metadata>
> >        <collection 
> > spec="/home/om/dods-data/thredds/roms/espresso/tmp/.*\.nc$"
> >                    name="espresso_2013_da_his_fmrc"
> >                    dateFormatMark="#espresso_his_#yyyyMMdd_HHmm"
> >                    olderThan="5 min" />
> >        <update startup="true" rescan="0 0/5 * * * ? *" trigger="allow" />
> >        <protoDataset choice="Penultimate">
> >           <netcdf 
> > xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2"; 
> > location="nc/cldc.mean.nc">
> >              <remove name="ocean_time" type="variable" />
> >           </netcdf>
> >        </protoDataset>
> >        <fmrcConfig regularize="false"
> >                    datasetTypes="TwoD Files Runs ConstantForecasts 
> > ConstantOffsets">
> >           <dataset name="Best" offsetsGreaterEqual="26" />
> >        </fmrcConfig>
> >     </featureCollection>
> >
> > -Lansing Madry
> >    Unidata
> >    Boulder, Colorado
> >
> 
> 

-Lansing Madry
  Unidata
  Boulder, Colorado

Ticket Details
===================
Ticket ID: KKM-141749
Department: Support THREDDS
Priority: Low
Status: Open