ANN: DataFramesMeta 0.9.0 release

I am proud to announce the 0.9.0 release of DataFramesMeta.jl.

We have three big changes inspired by DataFrameMacros.jl

  • Better support for row-wise transformations: We now export @rtransform, @rselect, and @rsubset, and @rorderby, which operate row-wise by default. You can still use @byrow to make some operations row-wise and others not.
  • To escape column names, you now use $ instead of the previous cols. i.e. @transform :y = f($x_str) will operate on the column whose name is represented by the variable x_str. You can also use $ on the LHS, @transform df $y_str = f(:x), and use it to refer to AsTable from DataFrames.jl, @transform df $AsTable = f(:x).
  • We add the @passmissing macro, for skipping missing values in row-wise operations. For example @rtransform df @passmissing :y = parse(Int, :x) will now create a missing value in the column :y when the column :x is missing, rather than throwing an error.

In my announcement of the 0.8.0 release, a user requested better documentation materials. So we re-wrote a popular dplyr tutorial to use DataFramesMeta.jl instead. It is now available in the documentation. I hope to port over data.table, Stata, Pandas tutorials as well and have them live in the documentation so tutorials will be easy to access.

Please download and use!



A small comment on a typo:

should be:

@rtransform df @passmissing :y = parse(Int, :x)