Handling NA in .csv files?

[ChristopherEeles]

We are having an issue where different csv writers (e.g., write.csv vs fwrite) have different default behaviours when writing NA values to disk.

As a result, the encoding of NAs has been changing in this repository and causing issues reading data for use in PharmacoDI.

To resolve this, we need to set standards around how NAs are encoded and perhaps adopt a default function to use when writing data from R and Python.