Merging multiple binary columns in R into one column preserving position

Question

I have have a data frame containing 30 columns in R that contain binary data. Every row contains exactly one 1. In other words the binary values are mutually exclusive. For all the thirty columns no two columns can contain 1 in the same row. Here is an illustration of what I mean.

Now obviously it is computationally enormously expensive to have this information distributed over thirty columns. What I want to do is merge all those 30 columns into one column that contains 30 different factor variables. For example the new column contains 2s in every row where the second column had 1s, 3s in every row where the third column had 1, etc. It is important that the original order is preserved and the positions do not get messed up as they act as index for other columns. So the above 3 columns would become this:

How can I achieve this in R?

Many thanks

you can use max.col(df)

talat
– talat

2017-04-04 12:00:38 +00:00
Commented Apr 4, 2017 at 12:00 — talat
– talat, Commented Apr 4, 2017 at 12:00

akrun · Accepted Answer · 2017-04-04 12:04:34Z

4

We can use max.col to find the index of the first value of the dataset

max.col(df1)
#[1] 1 2 3 1 3

Or with pmax

do.call(pmax, col(df1)*df1)
#[1] 1 2 3 1 3

data

df1  <- structure(list(v1 = c(1L, 0L, 0L, 1L, 0L), v2 = c(0L, 1L, 0L, 
0L, 0L), v3 = c(0L, 0L, 1L, 0L, 1L)), .Names = c("v1", "v2", 
"v3"), class = "data.frame", row.names = c(NA, -5L))

edited Apr 4, 2017 at 12:04

answered Apr 4, 2017 at 12:01

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

talat Over a year ago

The ties.method specification is unnecessary since there can only be a single 1 per row

akrun Over a year ago

@docendodiscimus Thanks, I came to know that you posted the comments first, but I was trying it for many times and I had to wait 60 secs until I can post one :-)

Sebastian · Accepted Answer · 2017-04-04 12:19:59Z

0

Thanks everyone. I also figured out a solution. If my data frame of 30 binary rows and 500 000 rows is called df, I just create a vector of 30 factors and loop through the factors:

factors = c(1:30)
newcol = rep(0, 500000)

for(f in factors){
    colvalues=df[,f]
    newcol[which(colvalues==1)]=f
}

answered Apr 4, 2017 at 12:19

Sebastian

4657 silver badges21 bronze badges

Collectives™ on Stack Overflow

Merging multiple binary columns in R into one column preserving position

2 Answers 2

data

2 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

data

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related