For missing (replace ismissing() with isnan() for NaNs) you can use also this trick, that makes all missing in numeric columns equal to 0, and all missing in strinng columns equal to "":
[df[ismissing.(df[!,i]), i] .= 0 for i in names(df) if Base.nonmissingtype(eltype(df[!,i])) <: Number]
[df[ismissing.(df[!,i]), i] .= "" for i in names(df) if Base.nonmissingtype(eltype(df[!,i])) <: String]