Last time my friend asked me about, how to split one columns in R dataframe to two columns. The example of data is looks like:

df <- structure(list(V1 = structure(c(2L, 1L, 3L), .Label = c("Andi", 
      "Bejo", "Rudi"), class = "factor"), V2 = structure(1:3, .Label = c("AA", 
      "AB", "TT"), class = "factor"), V3 = structure(1:3, .Label = c("AB", 
       "BB", "GG"), class = "factor"), V4 = structure(1:3, .Label = c("BB", 
        "CC", "CD"), class = "factor")), .Names = c("V1", "V2", "V3", 
        "V4"), class = "data.frame", row.names = c(NA, -3L))

Show example data:

##     V1 V2 V3 V4
## 1 Bejo AA AB BB
## 2 Andi AB BB CC
## 3 Rudi TT GG CD

Based on this dataframe we can use stupid way such as using “substr” the code looks like:

#first columns (the format is factor, so I convert it to characters)
a <- as.character(df$V1)
#second columns, we want to split this column to two columns, we use substr to get value
b <- substr(df$V2,1,1)
c <- substr(df$V2,2,2)

#third columns, we did samething with that we did for second columns
d <- substr(df$V3,1,1)
e <- substr(df$V3,2,2)

#forth columns, as well as previous columns, we split one columns to two columns. and we use substr for getting value
f <- substr(df$V4,1,1)
g <- substr(df$V4,2,2)

#Join vector (columns) to one matrix
new_matrix <- cbind(a,b,c,d,e,f,g)

After we create new matrix, we can convert it to dataframe, using code like this:

new_df <-
#to rename columns, you can use names(new_df), please look in previous tutorial if you did not know.

Now, we have new dataframe as we want.

##      a b c d e f g
## 1 Bejo A A A B B B
## 2 Andi A B B B C C
## 3 Rudi T T G G C D

The problem came when my friend said that he has 5000 columns. So, how do you think? you want to process 5000 columns data using that ways? it just look like stupid. After looking some information in Google, I found the best way. We can use library “splitstackshape”

First, Install and load that library:

## Warning: package 'splitstackshape' was built under R version 3.1.2
## Loading required package: data.table
## Warning: package 'data.table' was built under R version 3.1.2

Use this code:

new_df2 <- cSplit(df, 2:ncol(df), "", stripWhite = FALSE)

Just only using one line, now, you get dataframe as you want. see, your new dataframe.

##      V1 V2_1 V2_2 V3_1 V3_2 V4_1 V4_2
## 1: Bejo    A    A    A    B    B    B
## 2: Andi    A    B    B    B    C    C
## 3: Rudi    T    T    G    G    C    D

And now, you can use that syntax to 5000 columns in your dataframe. Oke, nice, see you in the next post :)