Recode one or more variables

Source: R/recodes.R

recodes recodes the values of one or more variables in a data frame

recodes(data, vars, from, to)

Arguments

data

a data frame.
vars

character vector of variable names.
from

a vector of values or conditions (see Details).
to

a vector of replacement values.

Value

a data frame

Details

  • For each variable in the vars parameter, values are checked against the list of values in the from vector. If a value matches, it is replaced with the corresponding entry in the to vector.

  • Once a given observation's value matches a from value, it is recoded. That particular observation will not be recoded again by that recodes() statement (i.e., no chaining).

  • One or more values in the from vector can be an expression, using the dollar sign ($) to represent the variable being recoded. If the expression evaluates to TRUE, the corresponding to value is returned.

  • If the number of values in the to vector is less than the from vector, the values are recycled. This lets you convert several values to a single outcome value (e.g., NA).

  • If the to values are numeric, the resulting recoded variable will be numeric. If the variable being recoded is a factor and the to values are character values, the resulting variable will remain a factor. If the variable being recoded is a character variable and the to values are character values, the resulting variable will remain a character variable.

Note

See the vignette for a worked example.

Examples

if (FALSE) {
# For variables A, B and C, convert 86, 99, and 999 to missing
df <- recodes(df, vars=c("A", "B", "C"), from=c(86, 99, 999), to=NA)

# For variables A, B, and C convert NA to 0
df <- recodes(df, vars=c("A", "B", "C"), from=NA, to=0)

# For variable X1, X2, and X3 convert 1 to Yes and 2 to No
df <- recodes(df, vars=c("X1", "X2", "X3"),
              from=c(1, 2), to=c("Yes", "No"))

# For variable SEX convert m to Male and f to Female
df <- recodes(df, vars="SEX",
              from=c("m", "f"), to=c("Male", "Female"))

# For variable OUTCOME convert Live to 0 and Die to 1
df <- recodes(df, vars="OUTCOME",
              from=c("live", "die"), to=c(0, 1))

# For variable AGE convert age <= 20 to young,
#    20 < age < =60 to middle, and age > 60 to old
df <- recodes(df, vars="AGE",
              from=c("$ < = 20", "$ > 20 & $ <= 60", "$  > 60"),
              to=c("young", "middle", "old"))

# Reverse code variable Rating from 1 to 10, to 10 to 1
df <- recodes(df, vars="Rating", from=1:10, to=10:1)

}