print and glimpse don't work on a tbl_df that contains data.frame column

[jennybc]

Since the new data_frame() performs no coercion of its arguments, there will likely be more data.frames (and tbl_dfs) running around with data.frame columns.

Here's a simple one and I've noticed that simple printing and glimpse()ing don't work.

suppressPackageStartupMessages(library(dplyr))

my_vec <- LETTERS[1:3]
my_df <- data.frame(month = month.abb[1:3],
                                        month_name = month.name[1:3])

dplyr_result <- data_frame(my_vec, my_df)
dplyr_result %>% str
dplyr_result
dplyr_result %>% print
dplyr_result %>% glimpse

devtools::session_info() %>%
    capture.output %>%
    grep("dplyr", ., value = TRUE)

And here's what it looks like in my Console:

> suppressPackageStartupMessages(library(dplyr))
> my_vec <- LETTERS[1:3]
> my_df <- data.frame(month = month.abb[1:3],
+ month_name = month.name[1:3])
> dplyr_result <- data_frame(my_vec, my_df)
> dplyr_result %>% str
Classes ‘tbl_df’, ‘tbl’ and 'data.frame':   3 obs. of  2 variables:
 $ my_vec: chr  "A" "B" "C"
 $ my_df :'data.frame': 3 obs. of  2 variables:
  ..$ month     : Factor w/ 3 levels "Feb","Jan","Mar": 2 1 3
  ..$ month_name: Factor w/ 3 levels "February","January",..: 2 1 3
> dplyr_result
Source: local data frame [3 x 2]

Error in `[.data.frame`(X[[2L]], ...) : undefined columns selected
> dplyr_result %>% print
Source: local data frame [3 x 2]

Error in `[.data.frame`(X[[2L]], ...) : undefined columns selected
> dplyr_result %>% glimpse
Variables:
Error in `[.data.frame`(X[[2L]], ...) : undefined columns selected
> devtools::session_info() %>%
+ capture.output %>%
+ grep("dplyr", ., value = TRUE)
Session info---------------------------------------------------------------------------
Packages-------------------------------------------------------------------------------
[1] " dplyr      * 0.3.0.9000 2014-11-06 Github (hadley/dplyr@8b0191a)   "

[hadley]

Here's an example of all the complex col types we need to support:

cdf <- data_frame(
  a = 1:3, 
  b = list(1, 3, 4), 
  c = data_frame(x = 1:3, y = 1:3), 
  d = matrix(1:6, ncol = 2)
)

[hadley]

Hmmmm, I'm now having second thoughts about supporting matrix and data frame columns, because it's not clear when they will be useful. I think I was getting confused with lists of data frames, which I can clearly see being used an as interim hierarchical form between normalised and unnormalised data. But data frame and matrix columns, just extend the columns in weird ways, and need a lot of special handling.

Data frames in particularly are problematic because the length of a data frame is the number of columns, where the length of a list-column is the number of rows.

cc @romainfrancois - I think this means we don't need special handling, and can happily throw an error if we see a data frame or matrix column. data_frame() will need a minor adjustment to use length() instead of NROW(). I'm not sure if we single them out for special handling errors as well.