c# - how to validate a file is Proper CSV or not in .net core?

Question

I have a APi which receives CSV files as IFormFile . I have to check if the sent file is a proper CS file or not. So i am doing below checks.

1. Checking the File Extension.
2. Checking the File content type.

Issue:- If any app will use the API, then it's feasible to change the file extension along with the content-type. So how to validate a proper CSV file? I didn't get any helpful article as of now.

e.g. a PDF file can be changed to a .CSV(in extension) file along with its content-type. But PDF file is not a valid CSV

NB:- Magic number is one of the process for .XLSX,.docx,.pdf etc.But for CSV its not applicable, tried the same & failed. Any other way to check it?

question from:https://stackoverflow.com/questions/65949566/how-to-validate-a-file-is-proper-csv-or-not-in-net-core

Answer 1

Closest you can get would be a robust TryParse method.

But instead of re-inventing the wheel, try first a few libraries, they might do the job:

https://github.com/TinyCsvParser/TinyCsvParser

https://github.com/nreco/csv

Note that CSV parsing can be a difficult task even though it's a simple format.

Even if you can't use a library, there are plenty of ideas you can grab from them.

If I were to detect CSV content, I'd do the following:

• ensure that a line contains readable characters
  • optionally detecting file enconding might help
• ensure that a line isn't incredibly long, else it's likely to be binary, see #1
• detect that first line has repeating separators
• try parse lines

More or less this:

• find 1st index of CR/LF or LF
• read up to that
• find separators in it
• try parse the rest of the file, check against column count
  • if it fails then it's probably not CSV

It's pretty much all the heuristics you can try unless I'm mistaken.