Transforming a simple data frame [duplicate]

Question

This question already has answers here:

Faster ways to calculate frequencies and cast from long to wide (4 answers)

How do I get a contingency table? (6 answers)

Closed 3 years ago.

I have a data frame that looks like this:

    ID    Club Type
1   1   arsenal 18
2   1   arsenal 31
3   1   arsenal 32
4   1   arsenal 14
5   2   chelsea 14
6   2   chelsea 31
7   2   chelsea 15
8   2   chelsea 32
9   3   fulham  27
10  3   fulham  31

I want to transform it into something like this:

    ID    Club  14 15 18 27 31 32
1   1   arsenal  1  0  1  0  1  1
2   2   chelsea  1  1  0  0  1  1
3   3   fulham   0  0  0  1  1  0

So in the transformed data frame, if a type matches with a particular club then it is given a 1 value and a 0 value otherwise.

My data frame is quite a bit bigger than this (otherwise I'd be doing them all manually).

Can someone give me a suggestion as to how I'd go about doing this? My problems stem from the fact that:

• types appear multiple times
• I'm not sure how I'd go about inserting binary values (I realise I could probably do something that returns True/False)
• the ID elements should collapse into one single element

Answer 1

library(reshape2) 

df <- read.table("clipboard",header=T)

dcast(df, ID + Club ~ Type, length)

  ID    Club 14 15 18 27 31 32
1  1 arsenal  1  0  1  0  1  1
2  2 chelsea  1  1  0  0  1  1
3  3  fulham  0  0  0  1  1  0 

Answer 2

The reshape alternative in R is not very pretty, but you can try:

x <- data.frame(table(mydf[-1]))          # Drop "ID" when tabulating
x$Freq <- as.numeric(as.logical(x$Freq))  # In case any are > 1
y <- reshape(x, direction = "wide", idvar="Club", timevar="Type")
merge(unique(mydf[1:2]), y)               # Reintroduce the ID
#      Club ID Freq.14 Freq.15 Freq.18 Freq.27 Freq.31 Freq.32
# 1 arsenal  1       1       0       1       0       1       1
# 2 chelsea  2       1       1       0       0       1       1
# 3  fulham  3       0       0       0       1       1       0