[R] Bug in levels() function?

[Peter Dalgaard]

Groot, Philip de wrote:
> Hello all,
>  
> I am not sure whether it actually is a bug, but it is not the behaviour I would expect. Please consider this:
>  
>   
>> Sibships
>>     
>  [1] Patient_2400 Patient_2400 Patient_345  Patient_345  Patient_8901
>  [6] Patient_8901 Patient_4008 Patient_4008 Patient_7991 Patient_7991
> [11] Patient_8353 Patient_8353 Patient_1212 Patient_1212 Patient_2168
> [16] Patient_2168 Patient_2760 Patient_2760 Patient_4726 Patient_4726
> [21] Patient_6699 Patient_6699 Patient_7641 Patient_7641 Patient_8263
> [26] Patient_8263 Patient_1389 Patient_1389 Patient_1618 Patient_1618
> [31] Patient_2410 Patient_2410 Patient_2612 Patient_2612 Patient_2721
> [36] Patient_2721 Patient_5053 Patient_5053 Patient_8458 Patient_8458
> [41] Patient_211  Patient_211  Patient_9004 Patient_9004 Patient_3423
> [46] Patient_3423 Patient_7413 Patient_7413 Patient_7815 Patient_7815
> [51] Patient_9232 Patient_9232 Patient_2267 Patient_2267 Patient_468
> [56] Patient_468
> 28 Levels: Patient_1212 Patient_1389 Patient_1618 Patient_211 ... Patient_9232
>  
>   
>> Comparison_Indices
>>     
>  [1]  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE
> [13] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
> [25] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
> [37] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE
> [49] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
>  
>   
>> Sibships[Comparison_Indices]
>>     
> [1] Patient_2400 Patient_2400 Patient_345  Patient_345  Patient_8901
> [6] Patient_8901 Patient_7413 Patient_7413
> 28 Levels: Patient_1212 Patient_1389 Patient_1618 Patient_211 ... Patient_9232
>
> The problem with this last command is that I would expect 4 levels (because only 8 "Comparison_Indices" are true, which is equal to 4 sibships. So: levels() does not take array indices into account or stated otherwise: if you use a subset in an array (vector), the levels() are not properly updated (to my opinion).
>  
> What I additionally found is the following:
>   
>> small_test <- factor(x=c("a", "b", "c"))
>> typeof(small_test)
>>     
> [1] "integer"
>
> The same happens to the Sibships that I defined as a factor? Why is it of type integer?
>  
> This is the version() output:
>   
>> version
>>     
>                _
> platform       x86_64-unknown-linux-gnu
> arch           x86_64
> os             linux-gnu
> system         x86_64, linux-gnu
> status
> major          2
> minor          6.1
> year           2007
> month          11
> day            26
> svn rev        43537
> language       R
> version.string R version 2.6.1 (2007-11-26)
>   
>  
> So: should I submit a Bug report?
>  
>   
No. This is all completely as designed. Factors are internally integers
(group codes), with a levels attribute that says what the codes mean. If
you want the full story, use dput(small_test) or class(small_test) or
str(small_test).

And subsetting a factor retains the original factor levels. To drop
unused levels, just use factor(f[index]) or f[index, drop=TRUE]. The
opposite behaviour can be even more annoying/dangerous because it leads
to empty cells dropping out of tables and bars disappearing from barplots.

-- 
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 35327907