Since I see that you've attached the screenshot, I assume that your data is not sensitive. For that reason I would ask you to put your question into reprex:
But to perhaps answer your question. You can always try to convert columns that you need to numeric with as.numeric command. It will replace all letters and such to NA's. You can then try the approach proposed by @Stephen.
And just to clarify a bit -- dplyr::select_if works on columns, not on rows. You can work on rows with, e.g., combination of dplyr::mutate and dplyr::case_when