att_gt() from did drops control group due to NA — how to include never-treated units?

I'm working with a staggered Difference-in-Differences (DiD) setup in R using the did package, where units are treated at different times and some are never treated (i.e., control group). Here's the code I'm using:

att_results <- att_gt(
    yname = "ln_Pris",
    tname = "Year",
    idname = "id",
    gname = "treatment_year",
    data = clean_data,
    panel = FALSE
)

Some units in my data are never treated, so their treatment_year is NA. When I run this, it seems like those units are being dropped from the estimation — I assume because gname is NA for the control group.

However, I need to include the never-treated units as the control group in the DiD analysis. I noticed there's a control_group argument in att_gt(), but I'm not sure how to use it in this situation.

Question:
How can I include never-treated units in the estimation? Do I need to manually assign something to gname for those units, or is setting control_group = "never treated" enough?

Thanks in advance!

Put treatment_year = 0 (instead of NA) for never-treated. You should take a close look at ?att_gt before using this function. Here is a screenshot from the documentation.