When creating a json extract the different column names can be defined. This would be great that it does not have to be defined manually but dynamically, I see 2 possibilities:
1/ getting the column list from another extract or transform
2/ have a new component type "HeaderUnion" that would unite the header and the data
this is comparable with Talend's DynamicSchema
this could be also useful for this purpose here: https://feedback.jedox.com/ideas/JS-I-1059
currently this is not possible to rearrange the order of output columns dynamically (the include pattern does not consider the order)
great, thanks Martin!
Thanks for the clarification on Jan 26. We have a better ideas of the request now and are discussing internally the feasibility of this.
any idea on this one ?
Let's precise a bit :
I have a json extract that extracts the following:
{"values":[["Country","2021","2022"],["FR",100,110],["DE",400,1100],["CH",150,140]]}
I iterate on $.values[1:] (to skip the first row)
the fields does not have any key, then I have to reference them as such:
[0]
[1]
[2]
then define aliases
But the first element of my json extract defines what is the header, if I could use this element to define the header of my extract, this could be more dynamic (e.g. for using in automatic normalizations)
that would be actually the same principle as the "Use header" option in csv extracts (meaning, using the first row as header)
hope this is clearer, if not, please don't hesitate to contact me!