group_by returns "Error: index out of bounds" when length of column name is g.t 39 characters

[elad-custora]

When we do:

library('dplyr')
df <- data.frame(a = 1:10, b = 1:10)
names(df) <- c('AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA', 'b')
df %>% group_by(AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA)

All is well. But:

names(df) <- c('AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA', 'b')
df %>% group_by(AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA)
Error: index out of bounds

[romainfrancois]

It happens here with lazyeval::auto_name here : https://github.com/hadley/dplyr/blob/master/R/group-by.r#L95

  new_groups <- lazyeval::auto_name(new_groups)

because:

> lazyeval::auto_name
function (x, max_width = 40)
{
    names(x) <- auto_names(x, max_width = max_width)
    x
}
<environment: namespace:lazyeval>

@hadley I understand the truncating is useful for when we are dealing with expression, but perhaps auto_name or auto_names could not truncate when it is a symbol. I'll start a new issue on lazyeval.

[romainfrancois]

Submitted a fix in lazyeval with which I now get:

> df <- data.frame(a = 1:10, b = 1:10)
> names(df) <- c('AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA', 'b')
> df %>% group_by(AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA)
Source: local data frame [10 x 2]
Groups: AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

   AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA  b
1                                        1  1
2                                        2  2
3                                        3  3
4                                        4  4
5                                        5  5
6                                        6  6
7                                        7  7
8                                        8  8
9                                        9  9
10                                      10 10
>
>
> names(df) <- c('AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA', 'b')
> df %>% group_by(AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA)
Source: local data frame [10 x 2]
Groups: AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

   AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA  b
1                                         1  1
2                                         2  2
3                                         3  3
4                                         4  4
5                                         5  5
6                                         6  6
7                                         7  7
8                                         8  8
9                                         9  9
10                                       10 10

[elad-custora]

Thank you!

[hadley]

I think I'd rather fix this in select_vars_. Let me think about it a little.

[dchudz]

This is pretty painful. Hoping for a fix soon.

[hadley]

@romainfrancois decided that your approach was the best after all, although I implemented it slightly differently.