rename_se

In seplyr: Improved Standard Evaluation Interfaces for Common Data Manipulation Tasks

seplyr is an attempt to work with the dplyr 0.7.* R package through standard evaluation interfaces with a minimum of change and minimum of cognitive friction. Beyond championing standard evaluation it attempts to introduce as few of its own opinions as possible.

A case in point is dplyr::rename()/seplyr::rename_se(). dplyr::rename() is used as follows.

suppressPackageStartupMessages(library("dplyr"))

datasets::mtcars %>%
  rename(cylinders = cyl, gears = gear) %>%
  head()

Notice dplyr::rename() renamings are written as assignments making them very similar to dplyr::mutate(). This is the reverse from how R usually handles list or mapping structures. In named vectors or named lists keys are written on the left and values are written on the right as follows.

mp <- c("cyl" = "cylinders", "gear" = "gears")
print(mp)

Because seplyr is intended to be a set of adapters for dplyr we simply adopt dplyr::rename()'s convention. This allows the user to mechanically translate their experience and expectations from scripting over dplyr::rename() to directly use seplyr::rename_se() as follows.

library("seplyr")

datasets::mtcars %.>%
  rename_se(., c("cylinders" := "cyl", "gears" := "gear")) %.>%
  head(.)

We hope this makes it easy to translate one-off analyses into re-usable scripts by incrementally replacing known variable names with parametric versions. The := operator is just a convenience function for building up maps, we could also have written rename_se(c("cylinders" = "cyl", "gears" = "gear")) or passed in a named vector built up elsewhere.

rename_se interprets all left-hand names as new column names and all right-hand names as old column names. This allows rename_se to be used to swap columns:

data.frame(a = 1, b = 2) %.>%
  rename_se(., c('a', 'b') := c('b', 'a'))

Please see help("%.>%", package="wrapr") for details on "dot pipe."

seplyr documentation built on Sept. 5, 2021, 5:12 p.m.