viewdf messes up display of frequencies if variable value starts at 0

[rangaro]

If a variable starts at the value 0, viewdf will always show "0" for the frequencies of that level and then shift all the other frequencies by one (i.e., the frequencies of "0" are displayed where the "1" is, the frequencies of "1" are displayed where the "2" is, and the frequencies of the last value are omitted).

For an example, see this variable that I recoded to start with 0

I can replicate this issue by merely recoding a variable so that the starting value is 0.

[strengejacke]

Do you have a reproducible example? This works for me:

set.seed(123)
d <- data.frame(
  x = factor(sample(0:3, 20, TRUE)),
  y = rnorm(20),
  z = factor(sample(letters[1:4], 20, TRUE))
)

sjPlot::view_df(d, show.frq = TRUE, show.prc = TRUE)

Data frame: d

ID	Name	Label	Values	Value Labels	Freq.	%
1	x			0 1 2 3	3 7 8 2	15.00 35.00 40.00 10.00
2	y		range: -2.0-1.8
3	z			a b c d	7 3 5 5	35.00 15.00 25.00 25.00

^{Created on 2024-11-25 with reprex v2.1.1}

Can you also try datawizard::data_codebook() |> insight::print_html()?

[rangaro]

I have a data file in XLSX format where this happens, but not an example with random data. The code you asked me to run only throws an error message: "Error in gt::tab_options(): ! table_font_size must be a single string, not a <tbl_df> object. Run rlang::last_trace() to see where the error occurred."

[strengejacke]

What if you convert the tibble (from readxl, I guess) into a data frame and then try the code again?

[rangaro]

If I convert the tibble into a data frame, then the issue remains.

[rangaro]

When I convert the variable from numeric to categorial with "as.factor", the issue vanishes. If I then convert it back to numeric and re-apply the labels, the issue comes back.

[rangaro]

When I convert the variable from numeric to categorial with "as.factor", the issue vanishes. If I then convert it back to numeric and re-apply the labels, the issue comes back.

[rangaro]

Try
set.seed(123)
d <- data.frame(
x = as.numeric(factor(sample(0:3, 20, TRUE))),
y = rnorm(20),
z = factor(sample(letters[1:4], 20, TRUE))
)
expss::val_lab(d[1]) = c("0" = 0,
"1" = 1,
"2" =2,
"3" = 3)
sjPlot::view_df(d, show.frq = TRUE, show.prc = TRUE)

[rangaro]

If you leave out the labeling, you can still see that there is an issue because the range of x is reported by sjPlot to be from 1 to 4 while it is actually from 0 to 3. I think that's where the issue is coming from.