R.rename?

Every now and then I wonder “should this kind of function be in Ramda?”. I just went over the docs for Python’s Pandas library (R-like DataFrames), trying to get a sense of parts it currently made more convenient. Among these was an object to rename columns in a dataframe. I couldn’t directly see a Ramda equivalent easier than some combination of pipe, toPairs, map, and fromPairs, so given that I’d propose a R.rename({ a: 'b', foo: 'bar' }, { a: 1, foo: 2 }) -> { b: 1, bar: 2 }. Pretty random, I know, especially since Pandas is by no means a direct competitor to Ramda users. Otherwise maybe fun for cookbook. Now to get corrected on a super-straightforward equivalent after all!

On a side-note, other bits that stood out to me from its cheat sheet as not having evident Ramda equivalents were NA handling (we’d probably end up deferring to Maybes?) / sparse data, stat-like functions e.g. binning into quartiles/percentiles, as well as maybe pivoting/unpivoting (which is kinda interesting too). Not a common comparison, I know, but that’s why I felt it’d be interesting to check. 😃

[nicohvi]

Not to revive an old issue, but if someone runs into the case rewriting one or few keys, and feel like creating a custom function for all fields to be included seems like overkill I solved it this way:

const renameKeys = keyMap => obj => {
  return R.reduce((acc, key) => {
    keyMap[key] ? acc[keyMap[key]] = obj[key]
    : acc[key] = obj[key]
    return acc;
  }, {}, R.keys(obj));
};

[chughes87]

I found that you could also make creative use of over and lens along with dissoc to do a rename:

const renameXZ = pipe(over(lens(prop('x'), assoc('z')), identity), dissoc('x'));
renameXZ({x: 1, y: 2}); // { y: 2, z: 1 }