Change to standard column names in params2csv

[tnatt]

PR to standardize the output csv from params2csv when run on an ensemble. Resolves #668

• Extract case metadata into standard column names ENSEMBLESET, ENSEMBLE, REAL and ITER (similar as csv_merge).
• Removed the old Realization and Iter columns.
• Changed to only dropping constant parameters columns

Note, it was considered keep and deprecate the Realization and Iter columns for a period. Still not sure if this is needed.. Discussed with @alifbe and @rnyb, and we all tend to lean towards dropping these columns.

• this tool is not expected to be very much in use, it is mainly used as an post-processing step
  • The ERT WORKFLOW functionality is very new (just a couple of months) and has not been added to the Drogon dataset yet. Hence few people have probably picked up this functionality.
• the change is not breaking any logic, and it should be an easy fix for the users.

Any thought on this @mferrera

[mferrera]

This seems reasonable to me. I think Alif or Roger should probably stamp an approval over myself.

[rnyb]

LGTM

[mferrera]

Please update the commit message before merging 👍

[tnatt]

As discussed with @alifbe and @rnyb:
I've extended the ERT example with an example showing how to extract parameters for all iterations in one go. In addition I've changed the filename for the ERT workflow file to lower case and added "wf_" as prefix to align with standards in Drogon 👍