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

Re: 20021126: converting old wmo metar data to gdbm? (fwd)


===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================

---------- Forwarded message ----------
Date: Wed, 04 Dec 2002 21:37:52 -0500
From: "[ISO-8859-1] Christian Pagé" <address@hidden>
To: Robb Kambic <address@hidden>
     decoders <address@hidden>
Subject: Re: 20021126: converting old wmo metar data to gdbm?

Thanks Robb, I will definitely do it bypassing ldm since I don't 
want to mess with my working ldm setup. Thank you very much for the 
help, I will do it as soon as I can (I am in a rush right now for 
other things).

Christian Pagé
UQAM

On Mercredi, novembre 27, 2002, at 02:40 , Robb Kambic wrote:

> Christian,
>
> If you don't want mess with making queues and the LDM, bypass it going
> directly to GDBM.  YOu should look at the man page for gdbm. You might
> have to make a c programs to do the gdbm work. I modified the 
> script, ie
>
>>>
>>> #!/local/bin/perl
>>> #
>     $fname = $ARGV[0] ;
>      $gdbm = gdbm_open ($fname ....
>>> # Now begin parsing file and decoding observations breaking on 
>>> cntrl C
>>> $/ = "\cC" ;
>>>
>>> while( <STDIN> ) {
>>>     # extract product ID
>>>     /(\w{4}\d{2} \w{4} \d{6}})/s
>>>     $prodID = $1 ; #you might want to change the id
>         gdbm_store ($gdbm, $prodID, $_, "0664") ;
>>> }
>
>     gdbm_close( $gdbm ) ;
>
> Robb...
>
>
> On Wed, 27 Nov 2002, [ISO-8859-1] Christian Pagé wrote:
>
>> Hi Robb,
>>
>> I understand your suggested way. But I have some more technical
>> concerns:
>>
>> I must create then an ldm queue to hold the products, and have an ldmd
>> process (and pqsurf) to take care of processing the bulletins.
>> Since I do ingest realtime data on that machine, which I don't want to
>> interfere, how can I set up a special queue, and a specific run 
>> of ldmd
>> to transfer all those bulletins without interfering the realtime data
>> flux?
>>
>> Thanks again for your help,
>>
>> On Wednesday, November 27, 2002, at 11:33 AM, Robb Kambic wrote:
>>
>>>>> To: address@hidden
>>>>> From: =?ISO-8859-1?Q?Christian_Pag=E9?= <address@hidden>
>>>>> Subject: converting old wmo metar data to gdbm?
>>>>> Organization: UCAR/Unidata
>>>>> Keywords: 200211011624.gA1GO6X00936
>>>>
>>>> Hi,
>>>>
>>>> Is there a way to convert old metar archived data from raw wmo 
>>>> format
>>>> to gdbm format?
>>>> In realtime, it uses pqsurf with a DBFILE action.
>>>
>>> Christian,
>>>
>>> You are asking a tough question.  It can be done with some work
>>> though.  I
>>> read your support question from July 31 about a similiar problem.
>>> Here's
>>> some background info that really important.  For pqsurf to work off a
>>> ldm
>>> queue it needs to have the bulletins have the same name as the 
>>> original
>>> raw bulletin name. Otherwise, pqsurf will not find them and not 
>>> perform
>>> any actions on them.  What this means is that you can't take an hours
>>> worth of bulletin and pqinsert into an ldm queue because they would
>>> have
>>> a wrong name and pqsurf will not act on it.
>>>
>>> Solution:
>>>
>>> You have to break up the hourly files into the original bulletins, ie
>>> starting with the ^A and ending with the ^C.
>>>
>>> ^A^M^M
>>> 981 ^M^M
>>> SAFR31 LFPW 311230^M^M
>>> LFJL 311230Z AUTO 26010KT 9999 SCT058 SCT070 30/15 Q1022=^M^M
>>> LFBG 311230Z 36006KT 320V040 CAVOK 31/16 Q1024=^M^M
>>> ^M^M
>>> ^C
>>>
>>> Using pqinsert, send it to the ldm with name "SAFR31 LFPW 311230"
>>>
>>> ie
>>>
>>> % pqinsert -vl - -s 999 -f IDS -p "SAFR31 LFPW 311230"
>>> <rawbulletinFile>
>>>
>>> where <rawbulletinFile> contains:
>>>
>>> ^A^M^M
>>> 981 ^M^M
>>> SAFR31 LFPW 311230^M^M
>>> LFJL 311230Z AUTO 26010KT 9999 SCT058 SCT070 30/15 Q1022=^M^M
>>> LFBG 311230Z 36006KT 320V040 CAVOK 31/16 Q1024=^M^M
>>> ^M^M
>>> ^C
>>>
>>> Code:
>>>
>>>
>>> One could make a perl script to do the work. Something like this:
>>> Script name bulletinInsert
>>>
>>> #!/local/bin/perl
>>> #
>>> # Now begin parsing file and decoding observations breaking on 
>>> cntrl C
>>> $/ = "\cC" ;
>>>
>>> while( <STDIN> ) {
>>>     # extract product ID
>>>     /(\w{4}\d{2} \w{4} \d{6}})/s
>>>     $prodID = $1 ;
>>>     open( OUT,">raw" ) ;
>>>     pring OUT $_ ;
>>>     close OUT ;
>>>     `pqinsert  -vl - -s 999 -f IDS -p "$prodID" raw` ;
>>>     unlink "raw" ;
>>> }
>>>
>>> To run:
>>>
>>> % bulletinInsert < <raw hourly files>
>>>
>>> I would extract one bullletin and test pqinsert first, then look 
>>> at the
>>> queue using pqcat to make sure the prodID is correct. Then use
>>> bulletinInsert and check queue again with pqcat.  I didn't debug the
>>> script, you might want to run it in debug mode to check it.
>>>
>>> % perl -d bulletinInsert < <raw hourly files>
>>>
>>> Hopes this works for you,
>>> Robb...
>>>
>>>
>>>>
>>>> Thanks,
>>>>
>>>> Christian Pagé
>>>> address@hidden
>>>> http://meteocentre.com/    http://meteoalerte.com/
>>>>
>>>> Etudiant au Doctorat en Sciences de l'environnement UQAM
>>>> +1 514 987 3000 ext. 2376
>>>>
>>>>
>>>> ------- End of Forwarded Message
>>>>
>>>>
>>>
>>> =======================================================================
>>> ========
>>> Robb Kambic                            Unidata Program Center
>>> Software Engineer III                          Univ. Corp for Atmospheric 
>>> Research
>>> address@hidden                 WWW: http://www.unidata.ucar.edu/
>>> =======================================================================
>>> ========
>>>
>>>
>>
>> Christian Pagé
>> address@hidden
>> http://meteocentre.com/    http://meteoalerte.com/
>>
>> Etudiant au Doctorat en Sciences de l'environnement UQAM
>> +1 514 987 3000 ext. 2376
>>
>>
>
> =========================================================================
> ======
> Robb Kambic                              Unidata Program Center
> Software Engineer III                    Univ. Corp for Atmospheric Research
> address@hidden                   WWW: http://www.unidata.ucar.edu/
> =========================================================================
> ======
>
>

Christian Pagé
address@hidden
http://meteocentre.com/        http://meteoalerte.com/
Etudiant au Doctorat +1 514 987 3000 ext. 2376
Sciences de l'Environnement UQAM