Error using n()

[rdboyes]

I came across this issue while writing some examples for TidierPlots. It may be that I'm just doing something wrong, but this code:

df_h = @chain penguins begin
    @group_by(species)
    @summarize(
        mean_bill = mean(bill_length_mm),
        se_bill = std(bill_length_mm) / sqrt(n())
    )
    @mutate(
        lower = mean_bill - 1.96 * se_bill,
        upper = mean_bill + 1.96 * se_bill
    )
end

Fails with error "No method matching nrow()". You can get this to run by separating the steps, like so:

df_h = @chain penguins begin
    @group_by(species)
    @summarize(
        mean_bill = mean(bill_length_mm),
        sd_bill = std(bill_length_mm),
        n = n()
    )
    @mutate(
        se_bill = sd_bill / sqrt(n)
    )
    @mutate(
        lower = mean_bill - 1.96 * se_bill,
        upper = mean_bill + 1.96 * se_bill
    )
end

But I don't think you would need to separate it out like this in dplyr

[kdpsingh]

I could be wrong, but my understanding when I wrote this function is that it wraps nrow => cols in DataFrames.jl and that nrow can only be used in this form and not within anonymous functions. Thus, this is a known limitation in the implementation of n() and isn't readily fixable unless there's a better way to translate this to DataFrames.jl.

See text, "Note! If the expression of the form x => y is passed then except for the special convenience form nrow => target_cols it is always interpreted as cols => function" from https://dataframes.juliadata.org/stable/lib/functions/#DataAPI.nrow.

[kdpsingh]

I thought I had documented this limitation somewhere but it looks like I hadn't. Could certainly add it to docs or think of a better way to implement.