[R] problems with which

[Michael Bedward]

Hi Nicola,

Although you have subsetted the data.frame, the factor variable will
still include the empty level.

For example, the following...

foo <- data.frame(f = factor(c("a", "b", "c", "b", "a")), val = 1:5)
foo2 <- foo[ foo$f != "a", ]
foo2
str(foo2)

Produces this output...

  f val
2 b   2
3 c   3
4 b   4

'data.frame':	3 obs. of  2 variables:
 $ f  : Factor w/ 3 levels "a","b","c": 2 3 2
 $ val: int  2 3 4

You can drop the unused factor level "a" like this...

foo2$f <- foo2$f[, drop=TRUE]
str(foo2)

'data.frame':	3 obs. of  2 variables:
 $ f  : Factor w/ 2 levels "b","c": 1 2 1
 $ val: int  2 3 4


Michael


On 15 August 2010 22:03, Nicola Spotorno <nicola.spotorno at isc.cnrs.fr> wrote:
> Hello,
> thanks for your replies, but the logic operator is '!='. If I call
> sentence_trial the dataframe shows no 'an'  datapoints  as I want but I
> still have problems with 'str' and 'interaction_plot'.
>
> Nicola
>