I don't understand this part of your post :
so, that aside, if you want the character strings interpreted as numeric values, then do this.
library(tidyverse)
a <- c("4", "3,4", "7")
b <- c("3,5", "3", "7,4")
df <- tibble(a,b)
(df2 <- mutate(df,
across(everything(),
~parse_number(str_replace_all(.x,",",".")))))
if you have further requirements for what to do with these numbers, please ask again with more clarity around that please.