Change data types using a list of data type names

user_5651991 0
Viewed 1557

What is an elegant way to change the data types of data frames columns from a list of data type names?

Here's an example (change_to_data_types function is what I'm looking for):

my_df <- iris
my_types <- c("factor", "character", "double", "logical", "character")
my_df <- my_df %>% change_to_data_types(my_types)

my_types has the same number of elements as the number of columns in my_df and the conversion is done in the same order.

This is an example of an 'inelegant' way

my_df$Sepal.Length <- my_df$Sepal.Length %>% as.factor()
my_df$Sepal.Width <- my_df$Sepal.Width %>% as.character()
#etc...
data.table dplyr dataframe r
4 Answers

An option would be

library(tidyverse)
my_df[] <- map2(my_df, str_c("as.", my_types), ~ get(.y)(.x))

Or in base R

my_df[] <- Map(function(x, y) get(y)(x), my_df, paste0("as.", my_types))

-checking the class again

sapply(my_df, class)
# Sepal.Length  Sepal.Width Petal.Length  Petal.Width      Species 
#    "factor"  "character"    "numeric"    "logical"  "character"
user_37322710

Having fun with match.fun:

my_df[] <- lapply(seq_along(names(my_df)),
                  function(i) match.fun(paste0("as.", my_types[ i ]))(my_df[[ i ]]))


sapply(my_df, class)
# Sepal.Length  Sepal.Width Petal.Length  Petal.Width      Species 
#     "factor"  "character"    "numeric"    "logical"  "character"
user_6800680

EDIT: To avoid having malformed factors due to the direct conversion of numerics to factors, we can do:

lapply(seq_along(names(my_df)),
                 function(x){
                   if(is.numeric(my_df[,x]) &
                      my_types[x] =="factor"){
                    as.factor(as.character(my_df[,x]))
                   } 
                   else{
                     as(my_df[,x],my_types[x])
                   }
                 }
                   )

ORIGINAL:

We can do:

sapply(seq_along(names(my_df)),
       function(x)  as(my_df[,x],my_types[x]))
user_103237980

We can use mapply to feed utils::as pairs of columns and types. This won't work for the factor columns, so those are handled separately.

fcols <- my_types == "factor"
my_df[!fcols] <- mapply(as, my_df[!fcols], my_types[!fcols], SIMPLIFY = FALSE)
my_df[fcols] <- lapply(my_df[fcols], factor)
user_49122290
Related