r - Column index with row-wise maximum value

Question

I have the following data:

  Newspaper    Month Year          Date      Topic1      Topic2     Topic3      Topic4      Topic5
1  Scotsman December 2005 December 2005 0.013749700 0.000127470 0.38575261 0.000127470 0.070778523
2  Scotsman December 2005 December 2005 0.000165017 0.000165017 0.05219433 0.004611941 0.000165017
3  Scotsman December 2005 December 2005 0.000356507 0.024344932 0.01135670 0.000356507 0.000356507
4  Scotsman December 2005 December 2005 0.000185186 0.000185186 0.10796924 0.044639345 0.106613401
5  Scotsman December 2005 December 2005 0.065869506 0.009775978 0.09610254 0.017584819 0.000103681
6  Scotsman December 2005 December 2005 0.000153257 0.000153257 0.11443001 0.000153257 0.046316677

I would like to create a separate variable that corresponds to the TopicN with higher percentage.

In the case of the first article (row), it would be 3. Any idea?

Answer 1

You could use max.col() on the topic columns. If df is the data, try

max.col(df[grepl("^Topic", names(df))])
# [1] 3 3 2 3 3 3

So to add a new column MaxPct, we can do

df$MaxPct <- max.col(df[grepl("^Topic", names(df))])