How should I perform missing imputation in DataFrame.jl? E.g., for a given DataFrame, how to turn all missings to 0. Thanks in advance!
1
votes
1 Answers
4
votes
use coalesce and broadcasting. So assuming your data frame is stored in df variable then just do:
df .= coalesce.(df, 0)
Now, if you wanted to perform this substitution only in selected columns then do:
@. df[!, cols] = coalesce(df[!, cols], 0)
where cols is a column selector.
An alternative way to achieve this is to use transform!:
transform!(df, cols .=> ByRow(x -> coalesce(x, 0)), renamecols=false)
where cols is your column selector. Use names(df) for cols is you want to do the imputation in all columns of the DataFrame.
This approach is a bit more verbose in this case, but it is more flexible in general.
