yjunechoe / selectmore

selectmore

{selectmore} provides extensions built on top of {tidyselect} for more flexible and powerful column subsetting patterns.

Installation

You can install the development version of selectmore from GitHub with:

# install.packages("devtools")
devtools::install_github("yjunechoe/selectmore")

Example

library(selectmore)
library(dplyr, warn.conflicts = FALSE)
df <- data.frame(
  x_1 = 1,
  x_2 = 2,
  y_2 = 3,
  y_1 = 4,
  y_3 = 5,
  x_3 = 6,
  nomatch = 0
)
df
#>   x_1 x_2 y_2 y_1 y_3 x_3 nomatch
#> 1   1   2   3   4   5   6       0

Baseline comparison with matches():

df %>%
  select(matches("(x|y)_(\\d)"))
#>   x_1 x_2 y_2 y_1 y_3 x_3
#> 1   1   2   3   4   5   6

Match and order with match_order():

# Order with priority on keeping the letters (1st group) close together
df %>%
  select(match_order("(x|y)_(\\d)", c(1, 2)))
#>   x_1 x_2 x_3 y_1 y_2 y_3
#> 1   1   2   6   4   3   5

# Similar to above, but sort letters in reverse order
df %>%
  select(match_order("(x|y)_(\\d)", c(-1, 2)))
#>   y_1 y_2 y_3 x_1 x_2 x_3
#> 1   4   3   5   1   2   6

# Order with priority on keeping the numbers (2nd group) close together
df %>%
  select(match_order("(x|y)_(\\d)", c(2, 1)))
#>   x_1 y_1 x_2 y_2 x_3 y_3
#> 1   1   4   2   3   6   5

# Same as above, the 2nd group is "slowest" because it has more categories
df %>%
  select(match_order("(x|y)_(\\d)", order_by = "slowest"))
#>   x_1 x_2 x_3 y_1 y_2 y_3
#> 1   1   2   6   4   3   5

Rename columns using {unglue} syntax with match_rename():

df %>% 
  select(
    match_order("(x|y)_(\\d)"), # Match columns on pattern and sorts by fastest
    everything()                # Grab the leftover columns (`nomatch`)
  ) %>% 
  # Rename matching columns using `unglue::unglue_sub()` syntax
  rename(match_rename("{letter}_{number}", list(toupper, ~ .x * 10)))
#>   X_10 Y_10 X_20 Y_20 X_30 Y_30 nomatch
#> 1    1    4    2    3    6    5       0