dummy_cols: Fast creation of dummy variables

In fastDummies: Fast Creation of Dummy (Binary) Columns and Rows from Categorical Variables

Fast creation of dummy variables

Description

Quickly create dummy (binary) columns from character and factor type columns in the inputted data (and numeric columns if specified.) This function is useful for statistical analysis when you want binary columns rather than character columns.

Usage

dummy_cols(
  .data,
  select_columns = NULL,
  remove_first_dummy = FALSE,
  remove_most_frequent_dummy = FALSE,
  ignore_na = FALSE,
  split = NULL,
  remove_selected_columns = FALSE,
  omit_colname_prefix = FALSE
)

Arguments

.data	An object with the data set you want to make dummy columns from.
select_columns	Vector of column names that you want to create dummy variables from. If NULL (default), uses all character and factor columns.
remove_first_dummy	Removes the first dummy of every variable such that only n-1 dummies remain. This avoids multicollinearity issues in models.
remove_most_frequent_dummy	Removes the most frequently observed category such that only n-1 dummies remain. If there is a tie for most frequent, will remove the first (by alphabetical order) category that is tied for most frequent.
ignore_na	If TRUE, ignores any NA values in the column. If FALSE (default), then it will make a dummy column for value_NA and give a 1 in any row which has a NA value.
split	A string to split a column when multiple categories are in the cell. For example, if a variable is Pets and the rows are "cat", "dog", and "turtle", each of these pets would become its own dummy column. If one row is "cat, dog", then a split value of "," this row would have a value of 1 for both the cat and dog dummy columns.
remove_selected_columns	If TRUE (not default), removes the columns used to generate the dummy columns.
omit_colname_prefix	If TRUE (not default) and 'length(select_columns) == 1', omit pre-pending the name of 'select_columns' to the names of the newly generated dummy columns

Value

A data.frame (or tibble or data.table, depending on input data type) with same number of rows as inputted data and original columns plus the newly created dummy columns.

See Also

dummy_rows For creating dummy rows

Other dummy functions: dummy_columns(), dummy_rows()

Examples

crime <- data.frame(city = c("SF", "SF", "NYC"),
    year = c(1990, 2000, 1990),
    crime = 1:3)
dummy_cols(crime)
# Include year column
dummy_cols(crime, select_columns = c("city", "year"))
# Remove first dummy for each pair of dummy columns made
dummy_cols(crime, select_columns = c("city", "year"),
    remove_first_dummy = TRUE)

fastDummies documentation built on Sept. 11, 2024, 8:07 p.m.