CODEWITHSUNDEEP
rr-factor
Arihant
Arihant

Reputation: 589

How not to drop factor levels in R cross tabulated tables?

I have a data frame in the following format and I want to get cross tabulation table:

 T1   T2
 NN  NF
 FF  FF
 NF  FF
 NN  NN
 NN  FF
 NF  NF
 NF  NF
 FF  FF

I can create a simple cross tabulation table by using the table() function

table(T1,T2) Which gives me the following output:

   FF FN  NF  NN
FF  2  0   0   0 
FN  0  0   0   0 
NF  1  0   2   0

However the factor level NN is dropped by default. Is there a way I can retain the NN level in the table. Which will table output look like:

   FF FN  NF  NN
FF  2  0   0   0 
FN  0  0   0   0 
NF  1  0   2   0
NN  0  0   0   0

edit: 

  dput(dataframe) gives: 

   structure(list(T1 = structure(c(3L, 1L, 2L, 3L, 3L, 2L), .Label = c("FF", 
   "NF", "NN"), class = "factor"), T2 = structure(c(2L, 1L, 1L, 
   3L, 1L, 2L), .Label = c("FF", "NF", "NN"), class = "factor")), .Names = c("T1", 
   "T2"), row.names = c(NA, 6L), class = "data.frame")

Upvotes: 2

Views: 1045

Answers (1)

gung - Reinstate Monica
gung - Reinstate Monica

Reputation: 11903

The problem is that "FN" isn't listed in your input data. You can get around this by calling levels() and informing R that it is one of your levels, it just is a level that does not have any observations in your dataset. For example, 

dframe <- read.table(text=" T1   T2
 NN  NF
 FF  FF
 NF  FF
 NN  NN
 NN  FF
 NF  NF
 NF  NF
 FF  FF", header=T)

levels(dframe$T1) <- c("FF", "NF", "NN", "FN")
levels(dframe$T2) <- c("FF", "NF", "NN", "FN")
with(dframe, table(T1,T2))
    T2
T1   FF NF NN FN
  FF  2  0  0  0
  NF  1  2  0  0
  NN  1  1  1  0
  FN  0  0  0  0

One special note: you are overwriting the original levels here, so you need to put the new, nonexistent, level last, otherwise R will call your "NF" level "FN" and lead to all sorts of problems.

Upvotes: 2

Related Questions