Repeated columns names erase each other for xParse about d3-dsv

[Fil]

I ♻️ my code into a notebook (and added "empty names" as well)
https://observablehq.com/@fil/csv-duplicate-names

[Fil]

Fixed in 8ab1ab8 ; thank you!

[Fil]

I think it's a good idea, and a possible implementation is given in #73

Note however that it would be a breaking change (people who already have some code running and this type of data expect it to continue working).

[mcnuttandrew]

I like your solution, but I don't know if it's worth issuing a breaking change. I think just including some stuff in the documentation would probably get most people through the hurdle of identifying this error

[Fil]

I don't know… The thing is that, when the data has this shape (and when you don't control it), it's currently quite difficult to manipulate: you have to load it as text, then fiddle with the first line, then dsv.parse… I've had to do this literally last week.
(Plus, we're going to issue a major version soon, so having a breaking change is not that problematic.)

[mcnuttandrew]

Oh i didn't know a major version was coming! This seems like a great approach then