NcML aggregation in TDS

[Fabien Goblet]

Hi,
I have problems with NcML aggregation in my TDS catalog.
I don't know how extracting date coordinates from a filename for a 
joinNew type.

Here is my catalog.xml :

<?xml version="1.0" encoding="UTF-8"?>
<catalog name="THREDDS Server Mercator Catalog"
         xmlns="http://www.unidata.ucar.edu/namespaces/thredds/InvCatalog/v1.0"
         xmlns:xlink="http://www.w3.org/1999/xlink">

   <service name="OpenDAP" serviceType="OpenDAP" base="/thredds/dodsC/"/>

   <dataset name="Aggregation Catalog" ID="psy3v1_med_agg" 
urlPath="psy3v1_med_agg">
     <serviceName>OpenDAP</serviceName>
     <dimension name="time" length="0"/>
     <variable name="time" type="int" shape="time">
       <attribute name="_CoordinateAxisType" value="Time"/>
       <attribute name="units" value="days since 1950-01-01 00:00"/>
     </variable>
     <netcdf xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2">
        <aggregation dimName="time" type="joinNew">
          <variableAgg name="temperature"/>
	 <variableAgg name="salinity"/>
	 <variableAgg name="u"/>
	 <variableAgg name="v"/>
	 <variableAgg name="kz"/>
	 <variableAgg name="ssh"/>
	 <variableAgg name="mld"/>
	 <variableAgg name="mlp"/>
	 <variableAgg name="taux"/>
	 <variableAgg name="tauy"/>
	 <variableAgg name="qtot"/>
	 <variableAgg name="emp"/>
	 <variableAgg name="qsr"/>
          <scan dateFormatMark="test_#yyyyMMdd" 
location="/data/mercator/developpement/goblet/data/psy3v1/med/best/" 
suffix=".nc" subdirs="false"/>
        </aggregation>
      </netcdf>
    </dataset>
</catalog>

I have 6 test files in the directory : from test_20070101.nc to 
test_20070105.nc

And I get an error from Ferret :

yes? use "http://opendap-dev:8880/thredds/dodsC/psy3v1_med_agg"

** netcdf error
time

yes? show data

  name     title                                                    I 
         J         K         L
  TIME 
...       ...       ...       ...
        (invalid coordinate axis)
  TEMPERATURE
           temperature                                     1:409 
1:187     1:40      1:6
  SALINITY 
salinity                                      1:409     1:187     1:40      1:6
  U        zonal 
velocity                                  1:409     1:187     1:40      1:6
  V        meridian 
velocity                              1:409     1:187     1:40      1:6
  KZ       vertical eddy 
diffusivity                     1:409     1:187     1:40      1:6
  SSH      sea surface 
height                         1:409     1:187     ...       1:6
  MLD      temperature ocean mixed layer 
t     1:409     1:187     ...       1:6
  MLP      density ocean mixed layer 
thick      1:409     1:187     ...       1:6
  TAUX     windstress eastward Tx componen  1:409     1:187     ...       1:6
  TAUY     windstress northward Ty compone  1:409     1:187     ...       1:6
  QTOT     total net heat 
flux                          1:409     1:187     ...       1:6
  EMP      water 
flux                                      1:409     1:187     ...       1:6
  QSR      surface downward solar heat 
flu       1:409     1:187     ...       1:6

yes? show grid temperature

GRID GFM3
  name                          axis              # 
pts           start                end
  LONGITUDE            LONGITUDE     409mr         6W(-6)               45E
  LATITUDE                LATITUDE        187 
i          30.023N              47.927N
  DEPTH                   DEPTH (m)        40 
i-               0                    4000
  TIME                       T                        6 
r               1                    6

Here is the "GetASCII" response (in OpenDAP) with the time checking :

Dataset {
     String time[time = 6];
} psy3v1_med_agg;
---------------------------------------------
time[6]
"2007-01-01T00:00:00Z", "2007-01-02T00:00:00Z", 
"2007-01-03T00:00:00Z", "2007-01-04T00:00:00Z", 
"2007-01-05T00:00:00Z", "2007-01-06T00:00:00Z"




I don't know why I got the Time like an invalid coordinate axis in Ferret.
However I don't know if the problem came from TDS or Ferret.
I will be glad if you can tell me how I can do this aggregation.

Thanks,
Fabien.

==============================================================================
To unsubscribe thredds, visit:
http://www.unidata.ucar.edu/mailing-list-delete-form.html
==============================================================================