Add rename() for DataFrame and LazyFrame

[etiennebacher]

Not sure about which args we should have here. I used existing and new as in the Rust implementation but the Python implementation uses a named list like list(old = "new") that I also like.

What do you think?

library(polars)

pl$
  DataFrame(mtcars)$
  rename(c("mpg", "hp"), c("miles_per_gallon", "horsepower"))
#> shape: (32, 11)
#> ┌──────────────────┬─────┬───────┬────────────┬───┬─────┬─────┬──────┬──────┐
#> │ miles_per_gallon ┆ cyl ┆ disp  ┆ horsepower ┆ … ┆ vs  ┆ am  ┆ gear ┆ carb │
#> │ ---              ┆ --- ┆ ---   ┆ ---        ┆   ┆ --- ┆ --- ┆ ---  ┆ ---  │
#> │ f64              ┆ f64 ┆ f64   ┆ f64        ┆   ┆ f64 ┆ f64 ┆ f64  ┆ f64  │
#> ╞══════════════════╪═════╪═══════╪════════════╪═══╪═════╪═════╪══════╪══════╡
#> │ 21.0             ┆ 6.0 ┆ 160.0 ┆ 110.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 4.0  ┆ 4.0  │
#> │ 21.0             ┆ 6.0 ┆ 160.0 ┆ 110.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 4.0  ┆ 4.0  │
#> │ 22.8             ┆ 4.0 ┆ 108.0 ┆ 93.0       ┆ … ┆ 1.0 ┆ 1.0 ┆ 4.0  ┆ 1.0  │
#> │ 21.4             ┆ 6.0 ┆ 258.0 ┆ 110.0      ┆ … ┆ 1.0 ┆ 0.0 ┆ 3.0  ┆ 1.0  │
#> │ …                ┆ …   ┆ …     ┆ …          ┆ … ┆ …   ┆ …   ┆ …    ┆ …    │
#> │ 15.8             ┆ 8.0 ┆ 351.0 ┆ 264.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0  ┆ 4.0  │
#> │ 19.7             ┆ 6.0 ┆ 145.0 ┆ 175.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0  ┆ 6.0  │
#> │ 15.0             ┆ 8.0 ┆ 301.0 ┆ 335.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0  ┆ 8.0  │
#> │ 21.4             ┆ 4.0 ┆ 121.0 ┆ 109.0      ┆ … ┆ 1.0 ┆ 1.0 ┆ 4.0  ┆ 2.0  │
#> └──────────────────┴─────┴───────┴────────────┴───┴─────┴─────┴──────┴──────┘

^{Created on 2023-06-10 with reprex v2.0.2}

[eitsupi]

Awesome! Looks nice.

[etiennebacher]

Actually I definitely prefer using a named list, it's much easier to see what is replaced by what when we want to rename a lot of variables:

library(polars)

pl$
  DataFrame(mtcars)$
  rename(
    list(
      mpg = "miles_per_gallon", 
      hp = "horsepower",
      gear = "Number of forward gears",
      carb = "Number of carburators"
    )
  )
#> shape: (32, 11)
#> ┌────────────────────┬─────┬───────┬────────────┬───┬─────┬─────┬────────────────────┬─────────────┐
#> │ miles_per_gallon   ┆ cyl ┆ disp  ┆ horsepower ┆ … ┆ vs  ┆ am  ┆ Number of forward  ┆ Number of   │
#> │ ---                ┆ --- ┆ ---   ┆ ---        ┆   ┆ --- ┆ --- ┆ gears              ┆ carburators │
#> │ f64                ┆ f64 ┆ f64   ┆ f64        ┆   ┆ f64 ┆ f64 ┆ ---                ┆ ---         │
#> │                    ┆     ┆       ┆            ┆   ┆     ┆     ┆ f64                ┆ f64         │
#> ╞════════════════════╪═════╪═══════╪════════════╪═══╪═════╪═════╪════════════════════╪═════════════╡
#> │ 21.0               ┆ 6.0 ┆ 160.0 ┆ 110.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 4.0                ┆ 4.0         │
#> │ 21.0               ┆ 6.0 ┆ 160.0 ┆ 110.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 4.0                ┆ 4.0         │
#> │ 22.8               ┆ 4.0 ┆ 108.0 ┆ 93.0       ┆ … ┆ 1.0 ┆ 1.0 ┆ 4.0                ┆ 1.0         │
#> │ 21.4               ┆ 6.0 ┆ 258.0 ┆ 110.0      ┆ … ┆ 1.0 ┆ 0.0 ┆ 3.0                ┆ 1.0         │
#> │ …                  ┆ …   ┆ …     ┆ …          ┆ … ┆ …   ┆ …   ┆ …                  ┆ …           │
#> │ 15.8               ┆ 8.0 ┆ 351.0 ┆ 264.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0                ┆ 4.0         │
#> │ 19.7               ┆ 6.0 ┆ 145.0 ┆ 175.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0                ┆ 6.0         │
#> │ 15.0               ┆ 8.0 ┆ 301.0 ┆ 335.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0                ┆ 8.0         │
#> │ 21.4               ┆ 4.0 ┆ 121.0 ┆ 109.0      ┆ … ┆ 1.0 ┆ 1.0 ┆ 4.0                ┆ 2.0         │
#> └────────────────────┴─────┴───────┴────────────┴───┴─────┴─────┴────────────────────┴─────────────┘

^{Created on 2023-06-10 with reprex v2.0.2}

[eitsupi]

Actually I definitely prefer using a named list, it's much easier to see what is replaced by what when we want to rename a lot of variables:

I agree with that, but I think the old = "new" style is confusing...
I think new = "old" is better.

[etiennebacher]

I prefer old = "new" but other packages like dplyr have this new = old so I suppose it's better for consistency to also have new = old

I'll change it

[etiennebacher]

library(polars)

pl$
  DataFrame(mtcars)$
  rename(
    miles_per_gallon = "mpg", 
    horsepower = "hp",
    `Number of forward gears` = "gear",
    `Number of carburators` = "carb"
  )
#> shape: (32, 11)
#> ┌────────────────────┬─────┬───────┬────────────┬───┬─────┬─────┬────────────────────┬─────────────┐
#> │ miles_per_gallon   ┆ cyl ┆ disp  ┆ horsepower ┆ … ┆ vs  ┆ am  ┆ Number of forward  ┆ Number of   │
#> │ ---                ┆ --- ┆ ---   ┆ ---        ┆   ┆ --- ┆ --- ┆ gears              ┆ carburators │
#> │ f64                ┆ f64 ┆ f64   ┆ f64        ┆   ┆ f64 ┆ f64 ┆ ---                ┆ ---         │
#> │                    ┆     ┆       ┆            ┆   ┆     ┆     ┆ f64                ┆ f64         │
#> ╞════════════════════╪═════╪═══════╪════════════╪═══╪═════╪═════╪════════════════════╪═════════════╡
#> │ 21.0               ┆ 6.0 ┆ 160.0 ┆ 110.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 4.0                ┆ 4.0         │
#> │ 21.0               ┆ 6.0 ┆ 160.0 ┆ 110.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 4.0                ┆ 4.0         │
#> │ 22.8               ┆ 4.0 ┆ 108.0 ┆ 93.0       ┆ … ┆ 1.0 ┆ 1.0 ┆ 4.0                ┆ 1.0         │
#> │ 21.4               ┆ 6.0 ┆ 258.0 ┆ 110.0      ┆ … ┆ 1.0 ┆ 0.0 ┆ 3.0                ┆ 1.0         │
#> │ …                  ┆ …   ┆ …     ┆ …          ┆ … ┆ …   ┆ …   ┆ …                  ┆ …           │
#> │ 15.8               ┆ 8.0 ┆ 351.0 ┆ 264.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0                ┆ 4.0         │
#> │ 19.7               ┆ 6.0 ┆ 145.0 ┆ 175.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0                ┆ 6.0         │
#> │ 15.0               ┆ 8.0 ┆ 301.0 ┆ 335.0      ┆ … ┆ 0.0 ┆ 1.0 ┆ 5.0                ┆ 8.0         │
#> │ 21.4               ┆ 4.0 ┆ 121.0 ┆ 109.0      ┆ … ┆ 1.0 ┆ 1.0 ┆ 4.0                ┆ 2.0         │
#> └────────────────────┴─────┴───────┴────────────┴───┴─────┴─────┴────────────────────┴─────────────┘

^{Created on 2023-06-10 with reprex v2.0.2}

[etiennebacher]

@eitsupi I'm just re-requesting your review since there were changes since you approved

[sorhawell]

r-polars has in general a syntax deviation from py-polars . Wherever in py-polars where first input is a list or a dict e.g. pl.DataFrame({"a":[1,2,3]}) then pl$DataFrame(a=1:3) and pl$DataFrame(list(a=1:3)) are both valid. Also e.g. df.select([pl.col("a")]) then df$select(pl$col("a")). If enable named_expr it is also possible to write df$select(new_a=pl$col("a")) instead of $alias("new"). But the ladder has not stabilized in py-polars due to wildcard last I checked.

[etiennebacher]

so @sorhawell are you fine with the syntax in the last post?

[sorhawell]

so @sorhawell are you fine with the syntax in the last post?

Oh yes I missed that :) sry! Looks great :)

I will take a quick look at code also

[sorhawell]

forgot to add something

[eitsupi]

Awesome! Thanks!