How to prevent tabulate function returning frequencies of variables that are not present in the vector?

I have the vector v1, with 26 elements, either being 'Live' or 'Non-live'.

When I use tabulate(v1), it returns

tabulate(v1)
           Value    Count   Percent
            cat1        0      0.00%
            cat2        0      0.00%
            Live        5     19.23%
        Non-live       21     80.77%
        

This affects crosstab also. How to prevent tabulate and crosstab functions taking non-existing variables in the vector in to account?

[Satwik]

Hi,

Could you please share the contents of the vector v1 ?

[Stephen23]

"How to prevent tabulate function returning frequencies of variables that are not present in the vector?"

Most likely they are present in your vector.

The example is easy to replicate by providing TABULATE with a categorical vector containing those four categories:

v1 = categorical(repelem(["Live","Non-live"],[5,21]),["cat1","cat2","Live","Non-live"])

v1 = 1x26 categorical array
  Columns 1 through 14

     Live      Live      Live      Live      Live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live 

  Columns 15 through 26

     Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live      Non-live 

tabulate(v1)

     Value    Count   Percent
      cat1        0      0.00%
      cat2        0      0.00%
      Live        5     19.23%
  Non-live       21     80.77%

"How to prevent tabulate and crosstab functions taking non-existing variables in the vector in to account?"

Most likely those categories do exist in the data. Therefore the simplest solution is to not create those categories in the vector.

Answer 1

[Walter Roberson]

The problem is that v1 was created as a categorical with additional categories beyond the ones populated.

I do not know of an efficient way to strip off the additional categories. One work-around is

v1cats = categories(v1);
v1stripped = categorical(v1cats(uint32(v1)));
tabulate(v1stripped)

You can adjust the uint32 according to the maximum number of categories the object holds. In this particular case you could use uint8