GroupBy node request


GroupBy node is possibly my favourite node, and one of the most powerful.

It does however, have a limitation which I find quite frustrating.

Its that the aggregation section cannot be used effectively on table structures which change (i.e. number of columns varies depending on input file into workflow).

If user doesnt what to apply specific aggreations to specific column names, Is there a possibility to implement under the Aggregation tab, some generic placeholders for handling different types of columns in which user selects an aggregation for this column type. i.e. So when new columns appear, a user selected aggregation is automatically applied,

i.e. For Double or Integer, it applies Mean,

and For String it applies First.



Hi Simon,

first of all thanks a lot for all your valuable feedback. The extension of the GroupBy node to handle columns based on data type is definitely on the agenda. The idea was to add an additional tab where the user can specify an aggregation method for a selected data type e.g. mean for double.  This option would then be applied to all columns which are not handled by the default options e.g. group or aggregation column. This feature will not gone make it for 2.9. However we will implement it as soon as possible. Please let us know if you have further ideas of how to improve the GroupBy node or any other node. We are always happy for constructive feedback.



Brilliant, that sounds perfect.