mutate_if applies fun to all columns if predicate returns all false

[jenswirf]

df1 <- data_frame(
  a = 1:10,
  b = runif(10),
  c = letters[1:10]
)

df2 <- data_frame(
  a = 1:10,
  b = runif(10),
  c = 10:1
)

# works as intended, c is converted to a factor
df1 %>% mutate_if(is.character, as.factor)

# all columns become factors
df2 %>% mutate_if(is.character, as.factor)

[cderv]

duplicates of #2009 and #1989 and already fixed in dev version by #2011 .

With dplyr 0.5.0.900, it is working ok.

devtools::dev_mode()
#> Dev mode: OFF
library(dplyr) # dev version 0.5.0.9000

df1 <- data_frame(
  a = 1:10,
  b = runif(10),
  c = letters[1:10]
)

df2 <- data_frame(
  a = 1:10,
  b = runif(10),
  c = 10:1
)

# works as intended, c is converted to a factor
glimpse(df1)
#> Observations: 10
#> Variables: 3
#> $ a <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
#> $ b <dbl> 0.32494503, 0.73370927, 0.90391393, 0.79685123, 0.09798207, ...
#> $ c <chr> "a", "b", "c", "d", "e", "f", "g", "h", "i", "j"
df1 %>% mutate_if(is.character, as.factor) %>% glimpse()
#> Observations: 10
#> Variables: 3
#> $ a <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
#> $ b <dbl> 0.32494503, 0.73370927, 0.90391393, 0.79685123, 0.09798207, ...
#> $ c <fctr> a, b, c, d, e, f, g, h, i, j

# all columns stays the same
glimpse(df2)
#> Observations: 10
#> Variables: 3
#> $ a <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
#> $ b <dbl> 0.67622567, 0.53826781, 0.31735237, 0.99455487, 0.20520285, ...
#> $ c <int> 10, 9, 8, 7, 6, 5, 4, 3, 2, 1
df2 %>% mutate_if(is.character, as.factor) %>% glimpse()
#> Observations: 10
#> Variables: 3
#> $ a <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
#> $ b <dbl> 0.67622567, 0.53826781, 0.31735237, 0.99455487, 0.20520285, ...
#> $ c <int> 10, 9, 8, 7, 6, 5, 4, 3, 2, 1

[hadley]

Fixed in #2011