List support

[holloszaboakos]

I love the library. It is already way better than plotly. However every time I load a data that is not stored in a dataframe or csv it is rearly painful to convert from a per row data class based structure to a column based one. It would be very useful if I could call it for any list and in each lambda of the DSL I could say how to map the record to the to the value relevant for the lambda.

For example if I have the data class:
data class C( val x: String, val y: Int, val z:String )

My data would be a List and I would use it like.
groupby{it.z} or
x{it.x} or
y{it.y} or
color{it.z}

I hope I could explain my idea well. Let's talk about it!

[AndreiKingsley]

Hi! Thank you ❤️ .
Yes, we have considered adding a row api (similar to what we already have in DataFrame - see row expressions), but there are a few interface and implementation issues. In particular, the lets-plot uses a system of markers - that is, pointers to the columns of the dataframe (using column names), and these pointers are used in many parts of the api; this would be more difficult to implement via row elements referencing. However, we are still thinking about the idea - moreover, a simple version of such api is quite realistic to implement..

[AndreiKingsley]

Now, the best solution for your case, as I see:

1. Cast your list to DataFrame using .toDataFrame()
2. Use KProp API in DSL (it's not supported everywhere in DSL yet, but that's easily fixable)