Requirements for getting dataset into CEDA

[n-a-t-e]

These are the requirements for getting a dataset into CEDA:

• hosted on working ERDDAP server. We will soon be using orderByCount to estimate file sizes, in this case we would require version >= 1.80, which is pretty old anyways.
• Tabledap only, no grids
• Trajectories are not currently supported
• Dataset variables - Time/Lat/long/depth are required. If datasets don't have depth we set them to 0. If there was a need to accommodate datasets without depth we could probably handle that.
• a CKAN record, which provides things like title, organization, and eovs
• For Timeseries data types that have two lat/longs (Eg a mooring with a nominal lat/long and also one that is measured) use preciseLat/preciseLon. See https://coastwatch.pfeg.noaa.gov/erddap/download/setupDatasetsXml.html#cdmTimeSeries

cdm_data_type

• cdm_data_type is an important variable in CEDA, and the cf_role attribute that goes with it
• cdm_data_type must be one of TimeSeries,TimeSeriesProfile,Profile,Point. Can't be a Trajectory/TrajectoryProfile/Other
• If you aren't sure which cdm_data_type to use, see:
  • https://coastwatch.pfeg.noaa.gov/erddap/download/setupDatasetsXml.html#EDDTableCdmTypes
  • https://cfconventions.org/Data/cf-conventions/cf-conventions-1.9/cf-conventions.html#discrete-sampling-geometries

[pramod-thupaki]

Adding link to issues with ONC data here https://github.com/HakaiInstitute/ceda/issues/43

[n-a-t-e]

See https://github.com/HakaiInstitute/cioos-data-explorer/wiki/ERDDAP-Requirements-for-Data-explorer-datasets