Mathematical operations for DataFrames

[nfoti]

Common math operations that work along dimensions of Arrays would be useful for DataFrames (and DataArrays as discussed in issue #325). The use case I envision is there are many numerical columns and some categorical columns, and say we want to compute the average of the numerical columns. It would be great to be able to say df["mean_col"] = mean(df[:,numeric_colinds], 2, skipna=true) without having to convert to a DataMatrix first. If someone tries to take the mean of something that doesn't make sense an error can be thrown.

If people think this is useful I'm willing to take a stab at it. I prefer the mean(df, dim, skipna) interface to rowmeans and colmeans to stay consistent with the rest of Julia but if others prefer the R names I can go along with that.

Thanks.

[garborg]

Closing as ensuing discussions have also landed on 'no' to treating DataFrames this way.