Do not edit "N/A", "NA", and similar entries **by default**. about csv.jl HOT 3 CLOSED

This is issue related to CSV.jl not DataFrames.jl so I transferred it.

The thing you request is handled by missingstring keyword argument. You can read about it here https://csv.juliadata.org/stable/reading.html#missingstring.

If this is clear and enough for you I propose to close this issue.

from csv.jl.

To be clear, I did NOT want anything to change in DataFrames.jl. In fact I was preemptively request you to not change the default behavior. (Not sure what's a good way to do that? Maybe I should have posted in discourse instead.) DataFrame.jl's default behavior right now is missingstring="". That is how it should be. In Pandas, the default behavior is the equivalent of missingstring=["NA", "NAN", "NULL"], which is very annoying.

I am not sure right now if importing and then exporting a csv using CSV.jl changes the data. If it does, then yes it would be an issue worth addressing.

I will close this issue.

(Also, I'm new to Julia and it's been so great to see your discussions, videos, and your work generally @bkamins . Thank you.)

from csv.jl.

Tested this csv file with CSV.jl and it seems to work as I would expect. It takes "" as missing or when there are two commas one after another. "NA" or NA and similar values get treated as strings.

string,value1,value2,value3,value4,value5,value6,mis1,mis2,num
a,1,1,1,1,1,1,,"",1.1
b,NA,"NA",N/A,"N/A",NULL,"NULL",,"",1.2

from csv.jl.