Feature proposal: adding possibility to order columns in nodes like column filter, groupby, pivotting...

This could simplify some workflows were column reorder node need to be systematically added.

Ideally, the possibility to rename directly the column from the filter or groupby node could also be useful.

1 Like

Hi @szawadski,

thanks for your proposal. So if I understood you correctly you would like to reorder/rename columns in the output of your node (e.g. Pivot) without inserting additional nodes everytime.

Do you usually sort or rename the columns that are produced as part of the node or also other columns?
Is their a pattern in your sorting (e.g. "I always like to have my pivoted columns BEFORE my pivot group (EVENT_MONTH here)?

I am just asking out of curiousity and to understand how we could best design a feature to do that…
Thanks in advance,
Best regards,

I would say that this is useful proposal. Just to provide a possibility to manually arrange fields. As for pivot node the close it to Excel pivot the better.

Thanks for your interest. My need is to freely rename the output of nodes like pivot, groupby or joiner. Typical use cases are:

  • two successive joins on fields of similar nature but with different meaning like address_origin and address_destination - node allows to add a suffix on duplicate names but could better to have more flexibility
  • in the case of long field names that are not valid anymore (max 30 char in Oracle) after they are suffixed, which means that suffix need to be shortened or rename node need to be added before the groupby
  • the aggregation are also generating complex field names that it would be nice to rename on the go…
    Feel free to ask more questions if not clear. I am so glad we have chosen a tool with such a dynamic team supporting it!