Reshape two columns of data with spread

[jstitlow]

Sorry, not a bug, just a suggestion.

This doesn't work using the spread function:
spread <- (df,Gentoype,D)
Error: duplicate identifiers for rows (1,2,3...etc)

But this does:
spread <- (df,Gentoype,D)

Is there a way to ignore the identifiers when the values are not correlated, and just fill the new columns directly?

maybe this:
spread <- (df,Genotype,D, Literal=TRUE)

[hadley]

Could you please rework your reproducible example to use the reprex package ? That makes it easier to see both the input and the output, formatted in such a way that I can easily re-run in a local session.

[markdly]

Here's a minimal reprex based on the original post and a possible workaround solution.

reprex based on original post

#==== minimal reprex for original post ====
suppressPackageStartupMessages(library(tidyverse))

df <- tribble(
  ~D, ~Genotype,
  1.2, "GFP",
  3.9, "GFP",
  5.5, "GFP",
  2.7, "WT",
  4.8, "WT")

# As expected, trying to spread as usual doesn't work as Genotype values are not unique
df %>% spread(Genotype, D)
#> Error: Duplicate identifiers for rows (1, 2, 3), (4, 5)

# Add rowname column before spreading enables spread to work
# But this is still not the desired output
df %>%
  rownames_to_column() %>%
  spread(Genotype, D)
#> # A tibble: 5 x 3
#>   rowname   GFP    WT
#> *   <chr> <dbl> <dbl>
#> 1       1   1.2    NA
#> 2       2   3.9    NA
#> 3       3   5.5    NA
#> 4       4    NA   2.7
#> 5       5    NA   4.8

Possible workaround solution
The desired output contains values for GFP and WT on the same row. Assuming the GFP and WT values are already ordered correctly, then this can be achieved by adding a rowid for each group and then spreading.

#==== Workaround solution for desired output ====
df %>% 
  group_by(Genotype) %>%
  mutate(group_row = 1:n()) %>%
  spread(Genotype, D)
#> # A tibble: 3 x 3
#>   group_row   GFP    WT
#> *     <int> <dbl> <dbl>
#> 1         1   1.2   2.7
#> 2         2   3.9   4.8
#> 3         3   5.5    NA

[hadley]

@markdly thanks for the reprex and workaround!