I have this sample dataframe
df <- data.frame(status=c('single', 'married', 'married', "fff"),
education=c(3456, 898989899, 23939004, NA),
income=c('34.5%', '88.456%', '92', NA))
df
I execute the following code to replace the decimal points with a comma in the character columns only
df1<-df %>% mutate(across(where(is.character), ~ str_replace_all(., "\\.", ",")))
The output df1 is OK. Next, I want to put dot ONLY to the values of the numerical columns. For example I want to have 3.456, 898.989.899, etc
But I have difficulty doing this. I tried
df2<- df1 %>% mutate(across(where(is.numeric) , ~ prettyNum(., big.mark=".") ))
but although it seems to work, I get a warning message.
Warning message:
There were 5 warnings in `mutate()`.
The first warning was:
ℹ In argument: `across(where(is.numeric), ~prettyNum(., big.mark = "."))`.
Caused by warning in `prettyNum()`:
! 'big.mark' and 'decimal.mark' are both '.', which could be confusing
ℹ Run `dplyr::last_dplyr_warnings()` to see the 4 remaining warnings.
Also, in df2, I no longer have numeric columns, as indicated by str(df2). All
columns are character columns which is something that I do not want, because NA becomes "NA", so in the excel instead of having blank cells, I will be having "NA" strings. I want to maintain the numerical class of the corresponding columns.
Is there a way to avoid these complications?
Note that my real dataset contains hundrends of such numeric and character columns and thousands of rows, so a more general code would help.
Ensure you indicate what the decimal mark should be:
df1 %>%
mutate(across(where(is.numeric), ~prettyNum(., big.mark=".",decimal.mark = ',')))
status education income
1 single 3.456 34,5%
2 married 898.989.899 88,456%
3 married 23.939.004 92
4 fff NA <NA>