[Go-essp-tech] syntax on the DRS ID

[Drach, Bob]

Hi Sylvia,

In that case it would make more sense to match on the other properties,
excluding product, time_frequency, and cmor_table.

Note there is a difference between publication-level datasets (the ID does
not include the variable name) and atomic dataset (includes variable name).

--Bob


On 2/1/11 10:03 AM, "Sylvia Murphy" <Sylvia.Murphy at noaa.gov> wrote:

> Hi Bob,
> The goal here is to connect a model metadata simulation instance to the data
> COLLECTION within ESG.  As such, we will need to parse this ID to exclude
> product, time_frequency, and cmor_table, which I believe are at the atomic
> dataset level.
> 
> Metafor reported today they plan to try and put this ID into the output of the
> questionnaire.  I am cc'ing them on this email so they can see the syntax and
> can chime in about the granularity of information.
> Thanks,
> Sylvia
> On Jan 28, 2011, at 12:42 PM, Robert S. Drach wrote:
> 
>> Hi Sylvia,
>> 
>> The DRS publication-level dataset ID - the IDs you will have available to
>> match against - is defined in the DRS document:
>> 
>> http://cmip-pcmdi.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf
>> 
>> See section 3.4: Publication-level dataset_id encoding. The form of the ID
>> is:
>> 
>> cmip5.product.institute.model.experiment.time_frequency.realm.cmor_table.ense
>> mble
>> 
>> Please note there are a few caveats at this point:
>> 
>> - Although the CMIP5 best practice recommendation is to follow the DRS
>> specification for dataset IDs, it is possible that the drs_id attribute may
>> differ, if the person doing the publisher configures esgpublish to do so. In
>> contrast the drs_id form is hardwired into the esgpublish code. The question
>> then is which one to match against if they differ. (I would suggest the
>> drs_id, but this remains open for discussion).
>> - Determination of the product field is relatively complicated in some cases.
>> In some cases you may need file-level information such as time ranges to
>> generate the product.
>> - Depending on  the  particular requirements for matching  metafor
>> information to  datasets, it might make sense to  just match individual
>> fields. All the above fields (product, institute, etc.) are defined
>> individually in the dataset metadata on the gateway.
>> 
>> --Bob
>> 
>> Sylvia Murphy wrote:
>>> Hi Karl,
>>> I have been talking to the metafor group about the DRS_ID you mentioned on
>>> the call this week.  They don't think they are going to have it output
>>> directly in their file but I may be able to put it together from various
>>> pieces.  I wanted to double check with you on what pieces are going into the
>>> ID since it may be different from my recollections.  At one point the syntax
>>> was going to be:
>>> 
>>> institution_model_simulationname_ripvalue + other stuff
>>> is that still valid?  The "other stuff" (e.g. variable names) will be more
>>> granularity than I'll need.
>>> If you have some examples of actual ids that you have seen generated by data
>>> providers that would help as well.
>>> 
>>> Thanks,
>>> Sylvia
>>> ***********************************
>>> Sylvia Murphy
>>> NESII/CIRES/NOAA Earth System Research Laboratory
>>> 325 Broadway, Boulder CO 80305
>>> Email: sylvia.murphy at noaa.gov <mailto:sylvia.murphy at noaa.gov>
>>> Phone: 303-497-7753
>>> Fax: 303-497-7649
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> ------------------------------------------------------------------------
>>> 
>>> _______________________________________________
>>> GO-ESSP-TECH mailing list
>>> GO-ESSP-TECH at ucar.edu
>>> http://mailman.ucar.edu/mailman/listinfo/go-essp-tech
>>>  
>> 
> 
> ***********************************
> Sylvia Murphy
> NESII/CIRES/NOAA Earth System Research Laboratory
> 325 Broadway, Boulder CO 80305
> Email: sylvia.murphy at noaa.gov
> Phone: 303-497-7753
> Fax: 303-497-7649
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> GO-ESSP-TECH mailing list
> GO-ESSP-TECH at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/go-essp-tech