Casting unique features in column to variable names and dummy coding original features into variables in R

Question

Having an issue with how to dummy code the following dataset.

Example data, lets say dataframe = mydata:

ID |     NAMES      |
-- | -------------- |
1  | 4444, 333, 456 |
2  | 333            |
3  | 456, 765       |

I'd like to cast only the unique variables in NAMES as column variables and code if each row has that variable or not i.e 1 or 0

Desired Output:

ID |     NAMES      | 4444 | 333 | 456 | 765 |
-- | -------------- |------|-----|-----|-----|
1  | 4444, 333, 456 |   1  |  1  |  1  |   0 |
2  | 333            |   0  |  1  |  0  |   0 |
3  | 456, 765       |   0  |  0  |  1  |   1 |

what I've done so far is created a vector of unique

split <- str_split(string = mydata$NAMES,pattern = ",")

vec <- unique(str_trim(unlist(split)))
remove <- ""
vec <- as.data.frame(vec[! vec %in% remove])
colnames(vec) <- "var"
vecRef <- as.vector(vec$var)

namesCast <- dcast(data = vec,formula = .~var)
namesCast <- nameCast[,2:ncol(namesCast)]

This yields a vector of unique NAMES with spaces/irregularities removed. From there I have no idea how to do the matching/dummy coding so any help would be greatly appreciated!

A5C1D2H2I1M1N2O1R2T1 A5C1D2H2I1M1N2O1R2T1 · Accepted Answer · 2014-12-03T15:10:55

You can use cSplit_e from my "splitstackshape" package, like this:

library(splitstackshape)
cSplit_e(mydata, "NAMES", sep = ",", type = "character", fill = 0)
#   ID          NAMES NAMES_333 NAMES_4444 NAMES_456 NAMES_765
# 1  1 4444, 333, 456         1          1         1         0
# 2  2            333         1          0         0         0
# 3  3       456, 765         0          0         1         1

If you want to see the underlying function that is called when you use those arguments, you can look at splitstackshape:::charMat, which takes a list generated by strsplit and creates a matrix from it.

Calling the function directly would give you something like this:

splitstackshape:::charMat(
  lapply(strsplit(as.character(mydata$NAMES), ","), 
         function(x) gsub("^\\s+|\\s$", "", x)))
#      333 4444 456 765
# [1,]   1    1   1  NA
# [2,]   1   NA  NA  NA
# [3,]  NA   NA   1   1

Casting unique features in column to variable names and dummy coding original features into variables in R

1 Answers