In IqisGnahz/rlookup: Lookup utilities for recoding strings in a data frame

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

• Use rlookup to create a value map from a data frame.
• Make desired change to the value map file.
• Apply the value map to the data frame.

Attach and load rlookup. The %>% piping operator from magrittr is strongly recommended.

library(rlookup)
library(magrittr)

Example data

We are using the good old mtcars data to illustrate the use of this package.

mtcars <- mtcars %>% tibble::rownames_to_column("model")
head(mtcars)

Make a value map

We desire to recode cyl, vs, am variables, which are respectively cylinder count, engine configuration, and transmission type.

A value map is called a lookup in rlookup; it is a data frame that has three columns: col_name, old_value and new_value, containing respectively:

• column name
• distinct values of the column, to be replaced with
• manually specified values for replacement

Use build_lookup() function to select desired columns from a data frame to make a value map. The new_value column is empty.

mtcars_lookup <- mtcars %>% build_lookup(cyl, vs, am)
mtcars_lookup

Write and read value map to file

Specify the path in write_lookup(). Give file name a csv suffix, as the output is in CSV format.

# Write the value map
# mtcars_lookup %>% write_lookup("mtcars_lookup.csv")

# Examine the file by reading it
# read_lookup("mtcars_lookup.csv")

Edit the value map

Edit the new_value column using a spreadsheet program, such as OpenOffice Calc or Microsoft Excel; if unavailable, use a free CSV editor such as {Ron's Editor}[https://www.ronsplace.eu/Products/RonsEditor].

From your R session, edit_lookup() is available to make a quick edit. If the funciton is supplied with a value map data frame, it will return the modified value map;If supplied with a file path to a value map file, modifications will be directly saved to the file. In both instances, the user will be asked to confirm the change.

You may delete value sets (rows) if you don't want to map them.

Do not try anything funny. As with many R programs, rlookup is not designed to protect the user from himself.

# Let's pretend that you've made desired change to the `new_value`.
read_lookup("mtcars_lookup_edited.csv") %>% write_lookup("mtcars_lookup.csv", overwrite = TRUE)
read_lookup("mtcars_lookup.csv")

Replace values

Now the value map is ready, finally it's time to replace values in the the data frame!

mtcars_recoded <- mtcars %>% use_lookup("mtcars_lookup.csv")
mtcars_recoded

Every value in their corresponding column has changed. That's basically it!

IqisGnahz/rlookup documentation built on May 9, 2019, 3:25 a.m.