"Missing" Values

[m-knopp]

Hi,

I have a hard time understanding why empty Strings are interpreted as missing by default. missing represents a value that exists, but we don't have access to. Why would we assume that with no semantic information about the data we are parsing?
"" is not a missing value, it is just an empty String and should be treated as such.

This is really awkward when you try further operations that fail because the type of a column is now Union{Missing, String}.
Also I found that
CSV.read("myfancytable.csv", DataFrame, missingstring="")
does not replace the "missing" values with empty Strings, they are still missing values.
CSV.read("myfancytable.csv", DataFrame, missingstring="abc")
does not replace "missing" values with String("abc"), but with nothing values.

My suggestion is to use String("") or nothing as the default value for an empty table cell.

[hhaensel]

You have two options:

• provide the option missingstring = nothing
CSV.read("myfancytable.csv", DataFrame, missingstring=nothing)
• provide an explicit type for the columns in question
  CSV.read("myfancytable.csv", DataFrame, types = Dict(:mystringcolumn => String))

[hhaensel]

I agree that it's a bit strange that

CSV.read("test.csv", DataFrame, missingstring = String[])

doesn't provide the same result as

CSV.read("test.csv", DataFrame, missingstring = nothing)