spreading data frame with nested columns

[cnjr2]

Spreading of data frames with multiple nested columns returns an error. (For more context on when one might encounter this issue see #197)

library("dplyr")
library("tidyr")
library("purrr")
library("broom")

data <- data_frame(
    name = c("Alex", "Alex", "Alex", "Tim", "Tim", "Tim"),
    year = c(1990, 1991, 1992, 1990, 1991, 1992),
    height = c(160, 165, 170, 120, 134, 150),
    weight = c(50, 52, 53, 48, 48, 52)
)

data
## name  year height weight
## (chr) (dbl)  (dbl)  (dbl)
## 1  Alex  1990    160     50
## 2  Alex  1991    165     52
## 3  Alex  1992    170     53
## 4   Tim  1990    120     48
## 5   Tim  1991    134     48
## 6   Tim  1992    150     52

# nest
data <- nest(data, year, height, weight)

data
## name           data
## (chr)          (chr)
## 1  Alex <tbl_df [3,3]>
## 2   Tim <tbl_df [3,3]>

Then we build a linear model where we only estimate the slope.

data <- data %>%
  mutate(
    model = map(.$data, ~lm(year ~ height + 0, data = .))
  )

data
## name           data     model
## (chr)          (chr)    (chr)
## 1  Alex <tbl_df [3,3]> <S3:lm>
## 2   Tim <tbl_df [3,3]> <S3:lm>

Then we extract model information with tidy and then unnest it:

# extract model statistics
tidy_model <- data %>%
  mutate(tidy = map(model, tidy)) %>%
  unnest(tidy)

tidy_model
## name           data   model   term estimate std.error statistic      p.value
## (chr)          (chr)   (chr)  (chr)    (dbl)     (dbl)     (dbl)        (dbl)
## 1  Alex <tbl_df [3,3]> <S3:lm> height 12.05941 0.2074858  58.12160 0.0002958912
## 2   Tim <tbl_df [3,3]> <S3:lm> height 14.66374 0.9394218  15.60932 0.0040791367

# then we gather
tidy_model <- tidy_model %>%
  gather(variable, value, estimate, std.error, statistic, p.value) %>%
  unite(variable, term, variable)

tidy_model
##   name           data   model         variable        value
##  (chr)          (chr)   (chr)            (chr)        (dbl)
## 1  Alex <tbl_df [3,3]> <S3:lm>  height_estimate 1.205941e+01
## 2   Tim <tbl_df [3,3]> <S3:lm>  height_estimate 1.466374e+01
## 3  Alex <tbl_df [3,3]> <S3:lm> height_std.error 2.074858e-01
## 4   Tim <tbl_df [3,3]> <S3:lm> height_std.error 9.394218e-01
## 5  Alex <tbl_df [3,3]> <S3:lm> height_statistic 5.812160e+01
## 6   Tim <tbl_df [3,3]> <S3:lm> height_statistic 1.560932e+01
## 7  Alex <tbl_df [3,3]> <S3:lm>   height_p.value 2.958912e-04
## 8   Tim <tbl_df [3,3]> <S3:lm>   height_p.value 4.079137e-03

# finally we want to spread again
tidy_model %>%
  spread(variable, value)

## Error in sort.int(x, na.last = na.last, decreasing = decreasing, ...) :
##   'x' must be atomic

Should this not return something like this:

tidy_model %>%
  spread(variable, value)

## name             data   model_A  model_B height_estimate height_p.value height_statistic height_std.error
## (chr)            (chr)    (chr)    (chr)         (dbl)          (dbl)            (dbl)            (dbl)
## 1  Alex <tbl_df [3,3]>  <S3:lm>  <S3:lm>      0.20000000   1.843174e-13     3.453933e+12     5.790501e-14
## 2   Tim <tbl_df [3,3]>  <S3:lm>  <S3:lm>      0.06656805   2.449142e-02     2.598076e+01     2.562205e-03

[hadley]

Minimal reprex:

df <- tibble::data_frame(x = c("a", "b"), y = 1:2, z = list(1:2, 3:5))
spread(df, x, y)

[hadley]

And tidyr is scheduled for release to CRAN on Friday :)

[cnjr2]

looking forward to it!