Unicode BOM messes up first property name

First off, thanks for the great library! Its API is a work of art. 👍

I ran into an issue when parsing CSV files fetched from the Github API with headers: true. The first column didn’t have the property name I expected due to an invisible character which turned out to be the UTF-8 BOM (U+FEFF). I fixed it in my code via:

response = response.replace(/^\ufeff/, "");

However, I thought you guys might want to fix it in the library at some point (and include other BOMs too, like the UTF-16 BOM etc). It’s debatable whether this should be fixed in the library, and I could argue it both ways, but thought I should bring it to your attention and let you decide. 😃

[kaligrafy]

It happens if we use Papa with nodejs also if the csv files contain BOM. It is not that trivial to remove BOM when using a readstream. Would it be possible to remove the DOM when parsing the header automatically?

[cweezy]

@LeaVerou I was running into this problem too, and I fixed the problem a different way. I am using javascript’s FileReader API and was using readAsBinaryString on a CSV file. Switching to readAsText for CSV files got rid of these BOMs.

Not sure if this would fix your issue, but figured it was worth sharing.