netcdf file content order

[ggalibert]

For ease of readability purpose, the order of the variables, their variable attributes, the global attributes visible in a dump should be consistent with the one specified in the .xml template.

For example at the moment, ncdump -h file.nc | less would give :

netcdf file {
dimensions:
        TIME = 28191 ;
variables:
        byte CNDC_quality_control(TIME) ;
                CNDC_quality_control:standard_name = "sea_water_electrical_conductivity status_flag" ;
                CNDC_quality_control:valid_max = 9b ;
                CNDC_quality_control:long_name = "quality flag for sea_water_electrical_conductivity" ;
                CNDC_quality_control:_FillValue = 99b ;
                CNDC_quality_control:flag_meaning = "No_QC_performed Good_data Probably_good_data Bad_data_that_are_potentially_correctable B
ad_data Value_changed Not_used Not_used Not_used Missing_value" ;
                CNDC_quality_control:quality_control_conventions = "IMOS standard set using the IODE flags" ;
                CNDC_quality_control:flag_values = 0b, 1b, 2b, 3b, 4b, 5b, 6b, 7b, 8b, 9b ;
                CNDC_quality_control:valid_min = 0b ;
                CNDC_quality_control:quality_control_set = 1 ;
        double TIME(TIME) ;
                TIME:axis = "T" ;
                TIME:standard_name = "time" ;
                TIME:valid_max = 90000. ;
                TIME:long_name = "time" ;
                TIME:valid_min = 0. ;
                TIME:units = "days since 1950-01-01 00:00:00 UTC" ;
                TIME:calendar = "gregorian" ;
etc...

while the .xml template says :

<variable>
      <name>TIME</name>
      <encoder>double</encoder>
      <dimensions>
        <dimension name="TIME"/>
      </dimensions>
      <attributes>
        <attribute name="standard_name" value="time"/>
        <attribute name="long_name" value="time"/>
        <attribute name="units" value="days since 1950-01-01 00:00:00 UTC"/>
        <attribute name="calendar" value="gregorian"/>
        <attribute name="axis" value="T"/>
        <attribute name="valid_min" value="0."/> <!-- this guy needs to be casted in the relevant type -->
        <attribute name="valid_max" value="90000."/> <!-- this guy needs to be casted in the relevant type -->
      </attributes>
    </variable>
    <variable>
      <name>LATITUDE</name>
      <encoder>double</encoder>
      <attributes>
        <attribute name="standard_name" value="latitude"/>
        <attribute name="long_name" value="latitude"/>
        <attribute name="units" value="degrees north"/>
        <attribute name="axis" value="Y"/>
        <attribute name="reference_datum" value="geographical coordinates, WGS84 projection"/>
        <attribute name="valid_min" value="-90."/> <!-- this guy needs to be casted in the relevant type -->
        <attribute name="valid_max" value="90."/> <!-- this guy needs to be casted in the relevant type -->
      </attributes>
    </variable>
etc...

So that what we should have is :

netcdf file {
dimensions:
        TIME = 28191 ;
variables:
        double TIME(TIME) ;
                TIME:standard_name = "time" ;
                TIME:long_name = "time" ;
                TIME:units = "days since 1950-01-01 00:00:00 UTC" ;
                TIME:calendar = "gregorian" ;
                TIME:axis = "T" ;
                TIME:valid_min = 0. ;
                TIME:valid_max = 90000. ;
        double LATITUDE ;
                LATITUDE:standard_name = "latitude" ;
                LATITUDE:long_name = "latitude" ;
                LATITUDE:units = "degrees_north" ;
                LATITUDE:axis = "Y" ;
                LATITUDE:reference_datum = "geographical coordinates, WGS84 projection" ;
                LATITUDE:valid_min = -90. ;
                LATITUDE:valid_max = 90. ;
etc...