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

[netCDFJava #HIP-291542]: Question on NcML aggregation

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: [netCDFJava #HIP-291542]: Question on NcML aggregation
  • Date: Fri, 05 Apr 2013 15:53:15 -0600
Li,

I think you want a union of the datasets, if they are all overlapping in 
time and space, rather than trying to join them via a new dimension.  
I'm not familiar with the data, though, so correct me if that doesn't 
make sense.

-Lansing

<?xml version="1.0" encoding="UTF-8"?>
<netcdf xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2";>
   <aggregation type="union">
     <netcdf location="cortadv4_row00_col00.nc"/>
     <netcdf location="cortadv4_row00_col01.nc"/>
     <netcdf location="cortadv4_row00_col02.nc"/>
     <netcdf location="cortadv4_row00_col03.nc"/>
   </aggregation>
</netcdf>


On 4/2/2013 8:45 AM, Yuanjie Li - NOAA Affiliate wrote:
> New Client Reply: Question on NcML aggregation
>
> Thanks Lansing!
>
> Could you please help me to check another data set?
> http://data.nodc.noaa.gov/thredds/catalog/cortad/Version4/catalog.html
>
> These are gridded Coral Reef Temperature Anomaly Database. Each file
> contains time series gridded data cube for a certain region. Row and
> columns numbers specify the region.
>
> All the data have the same time duration (1981/11 to 2010/12) and same time
> interval (once per week).
>
> Shall I use joinNew for this data set?
>
> Thanks a lot,
> Li
>
>
>
>
>
>
>
>
>
>
>
>
> address@hidden> wrote:
>
>> Li,
>>
>> Unfortunately, you probably won't be able to aggregate across buoys at
>> this point in time.  A feature collection such as what you are describing
>> is on our wish list, but isn't yet implemented in a reliable fashion.  The
>> NcML aggregations would fail because of holes in the data, as aggregations
>> really need homogenous data sets.
>>
>> One thing I am doing, though, is pulling down a bunch of your buoy data
>> files to use as a test set.  I'd guess this is a "sometime in the next
>> year" horizon project.
>>
>> Regards,
>>    Lansing
>>
>>> Hi Lansing,
>>>
>>> It would be great if we could set up different buoys in one NcML
>>> aggregation.
>>>
>>> Hi Matt,
>>> Do you have idea about how many buoys and station names are there from
>> all
>>> the files?
>>>
>>>
>>> Thanks,
>>> Li
>>>
>>>
>>>
>>>
>>>
>>>
>>> *
>>>
>>>
>>> *
>>> address@hidden> wrote:
>>>
>>>> Li,
>>>>
>>>> After looking at the files you are serving up, I think joinExisting is
>> the
>>>> way to go.  You'll need to set up different NcML aggregations for the
>>>> different buoys (DUO and SEF, it looks like).  Is there some other
>>>> information you thought users would be selecting?  I'm not entirely
>> sure
>>>> what you mean in your comments below on joinExisting.
>>>>
>>>> -Lansing
>>>>
>>>>> Thanks Lansing and Matt!
>>>>>
>>>>> I put the three ways of aggregated files under the test THREDDS
>> server.
>>>>> Here are few comments.
>>>>>
>> http://data.nodc.noaa.gov/thredds/catalog/testdata/20130328/catalog.html
>>>>> joinNew: There is "timeagg" in the OPeNDAP selection, but the actual
>>>>> temperature values are extracted from the first timeagg, no matter
>> which
>>>>> timeagg we choose.
>>>>>
>>>>> joinExisting: this one picks up all the values for all the joined
>> files.
>>>>> Though there is no list of station_name, plfatform1, instrument, for
>> the
>>>>> users to select.
>>>>>
>>>>> Union: this one picks up only the first file in the aggregation, and
>>>> there
>>>>> is no way to select the files after.
>>>>>
>>>>> The joinExisting is the way to go.
>>>>>
>>>>> Thanks,
>>>>> Li
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> address@hidden> wrote:
>>>>>
>>>>>> Matt,
>>>>>>
>>>>>> I set up an aggregation using four of the files.  At it's simplest,
>>>> this
>>>>>> works to join the datasets along the existing time dimension:
>>>>>>
>>>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>>>> <netcdf xmlns="
>> http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2
>>>> ">
>>>>>>    <aggregation dimName="time" type="joinExisting">
>>>>>>      <netcdf id="1" location="DOU001_000_20070530_20070906.nc"/>
>>>>>>      <netcdf id="2" location="DOU001_000_20070906_20071107.nc"/>
>>>>>>      <netcdf id="3" location="DOU001_000_20080807_20081204.nc"/>
>>>>>>      <netcdf id="4" location="DOU001_000_20100803_20101117.nc"/>
>>>>>>    </aggregation>
>>>>>> </netcdf>
>>>>>>
>>>>>> Since you already have a good dimension for aggregating (time),
>> then
>>>> you
>>>>>> probably don't need to worry about doing a joinNew.  You would use
>>>> that if
>>>>>> there were no time data in the files, and you wanted to add it
>> into the
>>>>>> dataset you were serving, for instance.
>>>>>>
>>>>>> Let me know if there is something else you are trying to do, I'm
>> happy
>>>> to
>>>>>> help.
>>>>>>
>>>>>> -Lansing
>>>>>>
>>>>>>> Lansing,
>>>>>>>
>>>>>>> When I set it up using the "joinExisting" on the time dimension
>> it
>>>> seems
>>>>>> to
>>>>>>> aggregate them. Below is the ncml:
>>>>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>>>>> <netcdf xmlns="
>>>> http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2";>
>>>>>>> <variable name="lat">
>>>>>>> <attribute name="standard_name" type="String" value="latitude"/>
>>>>>>> <attribute name="axis" type="string" value="Y"/>
>>>>>>> <attribute name="units" type="string" value="degrees_north"/>
>>>>>>> </variable>
>>>>>>> <variable name="lon">
>>>>>>> <attribute name="standard_name" type="string" value="longitude"/>
>>>>>>> <attribute name="axis" type="string" value="X"/>
>>>>>>> <attribute name="units" type="string" value="degrees_east"/>
>>>>>>> </variable>
>>>>>>> <variable name="time">
>>>>>>> <attribute name="standard_name" type="string" value="time"/>
>>>>>>> <attribute name="axis" type="string" value="T"/>
>>>>>>> <attribute name="units" type="string" value="seconds since
>> 1970-01-01
>>>>>>> 00:00:00"/>
>>>>>>> </variable>
>>>>>>> <aggregation dimName="time" type="joinExisting">
>>>>>>> <netcdf id="1" location="
>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20070530_20070906.nc
>>>>>>> "/>
>>>>>>> <netcdf id="2" location="
>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20070906_20071107.nc
>>>>>>> "/>
>>>>>>> <netcdf id="3" location="
>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20080807_20081204.nc
>>>>>>> "/>
>>>>>>> <netcdf id="4" location="
>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20100803_20101117.nc
>>>>>>> "/>
>>>>>>> <netcdf id="5" location="
>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20101117_20110318.nc
>>>>>>> "/>
>>>>>>> </aggregation>
>>>>>>> </netcdf>
>>>>>>>
>>>>>>> Is there a reason the joinExisting would work, but not the
>> joinNew? I
>>>>>>> thought joinNew would be able to aggregate anything, since your
>>>> creating
>>>>>> a
>>>>>>> new variable to aggregate them with.
>>>>>>>
>>>>>>> I apologize if this is something simple that I am just missing,
>> I am
>>>> just
>>>>>>> starting to get my bearings with this.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Matt
>>>>>>>
>>>>>>>
>>>>>>> address@hidden> wrote:
>>>>>>>
>>>>>>>> Lansing,
>>>>>>>>
>>>>>>>> Below are a few of the files I am trying to aggregate.
>>>>>>>>
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20070530_20070906.nc.html
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20070906_20071107.nc.html
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20080807_20081204.nc.html
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20100803_20101117.nc.html
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20101117_20110318.nc.html
>>>>>>>> Here is the ncml for the joinNew I've been fiddling around
>> with:
>>>>>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>>>>>> <netcdf xmlns="
>>>> http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2
>>>>>> ">
>>>>>>>>   <!-- <variable name="time" type="int">
>>>>>>>> <attribute name="units" type="string" value="seconds since
>>>> 1970-01-01
>>>>>>>> 00:00:00"/>
>>>>>>>>   <attribute name="_CoordinateAxisType" value="Time" />
>>>>>>>>    </variable>-->
>>>>>>>>    <aggregation dimName="timeagg" type="joinNew">
>>>>>>>>    <variableAgg name="A"/>
>>>>>>>> <netcdf id="1" location="
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20070530_20070906.nc
>>>>>>>> "/>
>>>>>>>> <netcdf id="2" location="
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20070906_20071107.nc
>>>>>>>> "/>
>>>>>>>> <netcdf id="3" location="
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20080807_20081204.nc
>>>>>>>> "/>
>>>>>>>> <netcdf id="4" location="
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20100803_20101117.nc
>>>>>>>> "/>
>>>>>>>> <netcdf id="5" location="
>>>>>>>>
>> http://data.nodc.noaa.gov/thredds/dodsC/nmsp/bml/DOU001_000_20101117_20110318.nc
>>>>>>>> "/>
>>>>>>>>    </aggregation>
>>>>>>>>   </netcdf>
>>>>>>>>
>>>>>>>> For some reason the Aggregation Variables are not being
>> listed...
>>>> Am I
>>>>>>>> completely missing something?
>>>>>>>> I truly appreciate your assistance in this matter. :)
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>> Matt
>>>>>>>>
>>>>>>>>
>>>>>>>> address@hidden> wrote:
>>>>>>>>
>>>>>>>>> Matt,
>>>>>>>>>
>>>>>>>>> If you send me a few of the files, I can set up the
>> aggregation
>>>> and
>>>>>> see
>>>>>>>>> if there are any pitfalls in the data.  It's not always
>>>>>> straightforward.
>>>>>>>>> -Lansing
>>>>>>>>>
>>>>>>>>>> Lansing,
>>>>>>>>>>
>>>>>>>>>> The times "should be" unified, but I should look into that.
>>>>>> Probably, to
>>>>>>>>>> avoid complications a JoinNew would suffice in this case.
>>>>>>>>>>
>>>>>>>>>> Matt
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Fri, Mar 29, 2013 at 12:58 PM, Unidata netCDF Java
>>>>>>>>>> address@hidden> wrote:
>>>>>>>>>>
>>>>>>>>>>> Li,
>>>>>>>>>>>
>>>>>>>>>>> Are the times in the data sets unified (i.e., all the
>> same),
>>>> or
>>>>>> are
>>>>>>>>> they
>>>>>>>>>>> interlaced?  In other words, if you were to set up a
>> single
>>>> time
>>>>>> axis
>>>>>>>>> for
>>>>>>>>>>> all of the buoys, would the data line up on the axis, or
>>>> would it
>>>>>> be
>>>>>>>>>>> scattered around?  If the times are all the same, then you
>>>> should
>>>>>> be
>>>>>>>>> able
>>>>>>>>>>> to do a simple Union as described here:
>>>>>>>>>>>
>>>>>>>>>>>
>> http://www.unidata.ucar.edu/software/netcdf/ncml/v2.2/Aggregation.html
>>>>>>>>>>> Otherwise, you will need to to a JoinNew, wherein you
>> declare
>>>> a
>>>>>> new
>>>>>>>>>>> dimension as the aggregation dimension.  This is
>> described on
>>>> the
>>>>>>>>> same page.
>>>>>>>>>>> -Lansing
>>>>>>>>>>>
>>>>>>>>>>>> Hi Lansing,
>>>>>>>>>>>>
>>>>>>>>>>>> I think there are group of buoys that are reporting data
>>>>>>>>> simultaneously.
>>>>>>>>>>>> The data sets should be overlapping in space.
>>>>>>>>>>>> I included Matt, who is the data officer for this data
>> set,
>>>> in
>>>>>> our
>>>>>>>>>>>> conversation.
>>>>>>>>>>>>
>>>>>>>>>>>> Hi Matt,
>>>>>>>>>>>> Please correct me if the description about the data is
>> not
>>>>>> right.
>>>>>>>>>>>> Thanks a lot,
>>>>>>>>>>>> Li
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> On Fri, Mar 29, 2013 at 10:40 AM, Unidata netCDF Java
>>>>>>>>>>>> address@hidden> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> Li,
>>>>>>>>>>>>>
>>>>>>>>>>>>> Are the data sets from the buoys overlapping in time,
>>>> space,
>>>>>> or
>>>>>>>>> both?
>>>>>>>>>>>>>   That is, do you have many buoys that are reporting
>> data
>>>>>>>>>>> simultaneously, or
>>>>>>>>>>>>> are the individual data set series generated by each
>> buoy
>>>>>>>>> temporally
>>>>>>>>>>>>> distinct?
>>>>>>>>>>>>>
>>>>>>>>>>>>> Also, are the nc files generated by the buoys, or have
>>>> they
>>>>>> been
>>>>>>>>>>> generated
>>>>>>>>>>>>> through some post-processing from a raw data set?
>>>>>>>>>>>>>
>>>>>>>>>>>>> If the files are not too large, feel free to send me
>> a few
>>>>>>>>>>> representative
>>>>>>>>>>>>> files to work with locally, and I will try to set up
>> an
>>>>>>>>> aggregation.
>>>>>>>>>>>>> Regards,
>>>>>>>>>>>>>    Lansing Madry
>>>>>>>>>>>>>    Unidata
>>>>>>>>>>>>>    Boulder, Colorado
>>>>>>>>>>>>>
>>>>>>>>>>>>>> Dear Sir,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I am trying to reach the experts about NcML
>> aggregation.
>>>>>>>>>>>>>> I have a group of time series buoy nc files. I
>> tried to
>>>>>>>>> aggregate
>>>>>>>>>>> them
>>>>>>>>>>>>>> by time, but failed since each of them holds a time
>>>> series
>>>>>>>>> itself.
>>>>>>>>>>>>>> It would be great if anyone could advise me on how
>>>> should I
>>>>>>>>> aggregate
>>>>>>>>>>>>>> these data.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks and Regards,
>>>>>>>>>>>>>> Li
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Ticket Details
>>>>>>>>>>>>> ===================
>>>>>>>>>>>>> Ticket ID: HIP-291542
>>>>>>>>>>>>> Department: Support netCDF Java
>>>>>>>>>>>>> Priority: Normal
>>>>>>>>>>>>> Status: Open
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Ticket Details
>>>>>>>>>>> ===================
>>>>>>>>>>> Ticket ID: HIP-291542
>>>>>>>>>>> Department: Support netCDF Java
>>>>>>>>>>> Priority: Normal
>>>>>>>>>>> Status: Open
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Mathew Biddle, Oceanographer
>>>>>>>>>> NOAA/NODC UMD/ESSIC/CICS E/OC1
>>>>>>>>>> 1315 East-West Hwy
>>>>>>>>>> Silver Spring, MD 20910-3282
>>>>>>>>>> Phone: (301) 713-3272 X163
>>>>>>>>>> Email: address@hidden
>>>>>>>>>> http://www.nodc.noaa.gov/
>>>>>>>>>> http://www.facebook.com/noaa.nodc
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Ticket Details
>>>>>>>>> ===================
>>>>>>>>> Ticket ID: HIP-291542
>>>>>>>>> Department: Support netCDF Java
>>>>>>>>> Priority: Normal
>>>>>>>>> Status: Open
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Mathew Biddle, Oceanographer
>>>>>>>> NOAA/NODC UMD/ESSIC/CICS E/OC1
>>>>>>>> 1315 East-West Hwy
>>>>>>>> Silver Spring, MD 20910-3282
>>>>>>>> Phone: (301) 713-3272 X163
>>>>>>>> Email: address@hidden
>>>>>>>> http://www.nodc.noaa.gov/
>>>>>>>> http://www.facebook.com/noaa.nodc
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Mathew Biddle, Oceanographer
>>>>>>> NOAA/NODC UMD/ESSIC/CICS E/OC1
>>>>>>> 1315 East-West Hwy
>>>>>>> Silver Spring, MD 20910-3282
>>>>>>> Phone: (301) 713-3272 X163
>>>>>>> Email: address@hidden
>>>>>>> http://www.nodc.noaa.gov/
>>>>>>> http://www.facebook.com/noaa.nodc
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> Ticket Details
>>>>>> ===================
>>>>>> Ticket ID: HIP-291542
>>>>>> Department: Support netCDF Java
>>>>>> Priority: Normal
>>>>>> Status: Open
>>>>>>
>>>>>>
>>>>>
>>>>
>>>> Ticket Details
>>>> ===================
>>>> Ticket ID: HIP-291542
>>>> Department: Support netCDF Java
>>>> Priority: Normal
>>>> Status: Open
>>>>
>>>>
>>>
>>
>> Ticket Details
>> ===================
>> Ticket ID: HIP-291542
>> Department: Support netCDF Java
>> Priority: Normal
>> Status: Open
>>
>>
>
>
> Ticket Details
> ===================
> Ticket ID: HIP-291542
> Department: Support netCDF Java
> Priority: Normal
> Status: Open
> Link:  
> https://www.unidata.ucar.edu/esupport/staff/index.php?_m=tickets&_a=viewticket&ticketid=21827



Ticket Details
===================
Ticket ID: HIP-291542
Department: Support netCDF Java
Priority: Normal
Status: Closed