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

Re: Re: Re: GFS ensemble data: name change affects ldmd.conf regex?



Christian,

I suggest using the minimum necessary for the request.:
 request         CONDUIT "(status|MT.nam|prod/gefs.*TMPK/850)"

(requests all status messages, all nam, and just 850mb TMPK for the
ensembles)

In pqact.conf:
CONDUIT prod/gefs.*(ge.*)\.t..z\.pgrb.*TMPK/850
        PIPE    decoders/dcgrib2 -d data/gempak/logs/dcgrib2_CONDUITens.log
                -e GEMTBL=/home/gempak/NAWIPS/gempak/tables
                data/gempak/model/ens/\1_YYYYMMDDHH_TMPK_850.gem

Remember that you have to restart the LDM when making changes to
ldmd.conf.
Running "ldmadmin pqactHUP" is sufficient when modifying pqact.conf
only.

Steve Chiswell
Unidata User Support



>         > CONDUIT
>         > prod/gefs\.(........)/(..)/pgrb2a/(ge.*)\.t..z\.pgrb2af(.*)
>         > !grib2/ncep/(.*)/(.*)/(.*)/TMPK/850 Pa PRES! 
>         >         PIPE    decoders/dcgrib2 -d
>         data/gempak/logs/dcgrib2_CONDUITens.log
>         >         -e GEMTBL=/home/gempak/NAWIPS/gempak/tables
>         >         data/gempak/model/ens/\3_YYYYMMDDHH_TMPK_850.gem


On Fri, 2006-10-13 at 14:00 +0200, Christian Pagé wrote:
> Hi Steve,
> 
> I don't get any dcgrib2_CONDUITens.log file anywhere... maybe the
> error is in the ldmd.conf file instead of pqact.conf?
> 
> Christian
> 
> 2006/10/9, Steve Chiswell < address@hidden>:
>         Christian,
>         
>         Check your dcgrib2_CONDUITens.log file.
>         If you have a log message such as "decode_grib2 -34" cannot
>         determine
>         parameter name for [0 0 0 1], then you need to add the
>         ensemble TMPK
>         pressure level variable to g2varswmo2.tbl.
>         
>         The ensemble template for temperature will be 001 in the 4th
>         column. You
>         may need to add to your $GEMTBL/grid/g2varswmo2.tbl file
>         (second line):
>         000 000 000 001 Temperature                      K 
>         TMPK             0  -9999.00
>         
>         I have a complete table if you need that for other variables.
>         
>         Steve Chiswell
>         Unidata User Support
>         
>         On Fri, 2006-10-06 at 21:45 +0200, Christian Pagé wrote:
>         > Hello Steve, 
>         >
>         > I tried your suggestions with some modifications, but I
>         don't receive
>         > any product yet. I am a little bit lost as I am not sure
>         where there
>         > is a problem in the regex.
>         >
>         > In ldmd.conf :
>         > request         CONDUIT "status|ST.opnl/MT.nam|
>         prod/gefs.*TMPK/850 Pa PRES"
>         >
>         > In pqact.conf:
>         > CONDUIT
>         > prod/gefs\.(........)/(..)/pgrb2a/(ge.*)\.t..z\.pgrb2af(.*)
>         > !grib2/ncep/(.*)/(.*)/(.*)/TMPK/850 Pa PRES! 
>         >         PIPE    decoders/dcgrib2 -d
>         data/gempak/logs/dcgrib2_CONDUITens.log
>         >         -e GEMTBL=/home/gempak/NAWIPS/gempak/tables
>         >         data/gempak/model/ens/\3_YYYYMMDDHH_TMPK_850.gem
>         >
>         > But I have nothing being decoded and no logs neither...
>         Would you have
>         > an idea where is the error?
>         >
>         > Thanks!
>         >
>         > Christian
>         >
>         > 2006/10/3, Steve Chiswell <address@hidden>:
>         > > Christian,
>         > >
>         > > The 850 mb data will have a pattern such as:
>         > > prod/gefs.*TMPK/850 Pa PRES
>         > >
>         > > (the Pa is because the vertical coordinate of the grid
>         data is actually 
>         > > Pascals
>         > > though the 850 is scaled be 10-2 for printing equvalent
>         mb)
>         > >
>         > > You can see the names of each field in the status file,
>         such as:
>         > >
>         data/nccf/com/gens/prod/gefs.20061002/18/pgrb2a/gec00.t18z.pgrb2aanl 
> ! 
>         > > grib2/ncep/SPEC62MRF/#000/200610021800F000/TMPK/850 Pa
>         PRES! 000009
>         > >
>         > > Steve Chiswell
>         > > Unidata User Support
>         > >
>         > > On Fri, 2006-09-29 at 22:37 +0200, Christian Pagé wrote: 
>         > > > Steve,
>         > > >
>         > > > Thanks again for being so quick to help us!
>         > > > Since our bandwidth is a little bit limited, is it
>         possible, by
>         > > > example, to just ask for TMPK at 850 mb for ensemble
>         data? 
>         > > >
>         > > > Let's say that I have this pqact.conf action:
>         > > > CONDUIT
>         > > > prod/gefs\.(........)/(..)/pgrb2a/(ge.*)\.t..z
>         \.pgrb2af(.*)
>         > > > !grib2/ncep/(.*)/\ 
>         > > > (.*)/(.*)/TMPK/850 mb!
>         > > >
>         > > > Can I put this in the ldmd.conf to ask only for this
>         field??
>         > > > request      CONDUIT      "status|
>         ST.opnl/MT.nam|(prod/gefs.* 
>         > > > !grib2/ncep/.*/.*/.*/TMPK/850 mb!)"
>         > > >
>         > > > Regards,
>         > > >
>         > > > Christian
>         > > > UQAM
>         > > >
>         > > > 2006/9/29, Steve Chiswell < address@hidden>:
>         > > > > Christian,
>         > > > >
>         > > > > For the CONDUIT files on the idd, the NAM files will
>         match with
>         > > > > " MT.nam", and the new ensemble files will match with
>         "prod/gefs". The
>         > > > > other parts of your request line shown below will have
>         no effect since
>         > > > > there are no longer MT.eta headers in the data
>         > > > > stream, and the bmoav and cy23r strings don't match
>         any CONDUIT
>         > > > > products.
>         > > > >
>         > > > > A single request line for nam and ensembles (which
>         will receive the 
>         > > > > status files) can look like:
>         > > > >
>         > > > > request      CONDUIT      "ST.opnl/MT.nam|prod/gefs"
>         > > > >
>         > > > > If you want all the status files for the other
>         products too, but just 
>         > > > > the NAM and Ensemble grids, then
>         > > > >
>         > > > > request      CONDUIT      "status|ST.opnl/MT.nam|
>         prod/gefs"
>         > > > >
>         > > > >
>         > > > > Steve Chiswell
>         > > > > Unidata User Support
>         > > > >
>         > > > >
>         > > > > On Fri, 2006-09-29 at 22:10 +0200, Christian Pagé
>         wrote:
>         > > > > > Hi everyone, 
>         > > > > >
>         > > > > > WIth the new name change for GFS ensemble data in
>         the CONDUIT feed:
>         > > > > > how this affects the regex used in ldmd.conf feed
>         configuration?
>         > > > > >
>         > > > > > I currently have:
>         > > > > > request         CONDUIT "(status)|(ST.opnl/MT.(eta|
>         nam|bmoav|cy23r|ensg))"
>         > > > > >
>         > > > > > How must I change this request to take into account
>         of the name changes? 
>         > > > > > I am not sure how I should update this request line.
>         > > > > >
>         > > > > > Regards,
>         > > > > >
>         > > > > --
>         > > > > Steve Chiswell < address@hidden>
>         > > > > Unidata
>         > > > >
>         > > --
>         > > Steve Chiswell <address@hidden>
>         > > Unidata
>         > >
>         --
>         Steve Chiswell <address@hidden>
>         Unidata
> 
> 
-- 
Steve Chiswell <address@hidden>
Unidata