Mapreduce dataframe columns elementwise AND

General Usage
1

Is there a specific reason why I need to define myop in the snippet below?
Otherwise things do not work for me.

using DataFrames 

dff=DataFrame(a=trues(4),b=trues(4),c=2*ones(4))
cnms = [:a, :b]

#fails
mapreduce(x->dff[!,x], .& , cnms)  

#works
myop(x,y) = x .& y
mapreduce(x->dff[!,x], myop , cnms)
1 Like
3

I think this is a parsing thing. .& isn’t a function.

I’ve wanted this before as well, an easy way to refer to the broadcasted version of a function via, for example, Broadcast(fun).

also, note that you can use eachcol here instead of iterating through the names.

4

Hm. Ok.
regarding eachcol: in my actual use case I only iterate over a subset of all columns.