Suggestion:: xtab function similar to frq() about sjmisc HOT 8 CLOSED

For the first issue, see sjmisc::flat_table().

from sjmisc.

Thanks, but I already tried that earlier without success. Could it be that flat_table() only shows values for which there exists a label and omits all other values from the output?

from sjmisc.

Yes, you're right. You have to wrap the data frame in fill_labels():

x <- sample(1:4, 20, T)
y <- sample(1:4, 20, T)
set_labels(x) <- c(`first` = 1, `third` = 3, `fourth` = 4)
set_labels(y) <- c(`second` = 2, `fourth` = 4)
tib <- tibble::tibble(x, y)
flat_table(fill_labels(tib), x, y)

else, use to_label():

tib <- to_label(tib, add.non.labelled = T)
ftable(tib$x, tib$y)

from sjmisc.

I think it doesn't make sense to exclude non-labelled values, so I fixed this.

from sjmisc.

Thank you for the update! What do you think about showing variable labels for frq() and fill_table()? In my opinion this would make them basically "swiss army knives" so that one can have a brief look at all important information about variables with one single (non-html) command. If you want I can also open another issue for this.

By the way: I am just in the process of converting a Data Analysis course for Social Scientists from Stata to R and thanks to your package this works out incredibly well! :)

from sjmisc.

For frq(), variable labels are already shown (in the current GitHub-build). For flat_table(), I have no good idea where to show the labels, because you can have a table with 2, 3, 4 or more variables. And the object itself is an ftable(), so I can't easily format/manipulate the console-output.

By the way: I am just in the process of converting a Data Analysis course for Social Scientists from Stata to R and thanks to your package this works out incredibly well! :)

Glad to hear that, and thanks for the feedbacks you provide - this helps me improving the package not only for other users, but also for my purposes.

from sjmisc.

I see, thanks for the frq() update. There is one more thing I would like to suggest: I think frq() should either not display value labels for which no observations are occurent at all, or have a parameter which would enable this behavior. The background is the following: In several datasets I work with there are much more value labels assigned than actually observed later on. Using frq() on such variables spams the screen with a lot of unnecessary information. I am not sure whether this change is relevant for SPSS datasets, but it definitely is for Stata.

Edit: This would be equivalent to skip.zero from sjt.frq() right?

from sjmisc.

As a First workaround you could wrap the variable in drop_labels().

from sjmisc.