Feature Request: Function to parse only at runtime specified fields

[Adrian Sieber]

I'm looking for a easy way to parse just the at runtime provided fields. So something like:

parseFields :
    { fieldSeparator : Char }
    -> List String
    -> String
    -> Result Problem (List (List String))
    
parseFields ',' ["id", "name", "color"] "id,name,color\n1,bike shed,green\n"

If the specified fields are not available it should it result in an error (maybe there should also be a flag wether to allow additional fields in the CSV).

Maybe this is already possible with decodeCsv, but if so I'm not sure how, since into needs a function with a pre defined arity.

[Brian Hicks]

Can you say more about what you're trying to do? https://package.elm-lang.org/packages/BrianHicks/elm-csv/latest/Csv-Parser#parse may suit you but I'd like to know more anyway so we can keep you in type-safety land.

[Adrian Sieber]

I'm building a website where a user can specify which fields a CSV should have, and then validate CSVs against that schema. I.e. the available fields must be specified at runtime.

[Brian Hicks]

Ah, I see! When you say a schema, are you talking about the types of data in the columns or only their names? I can't think of a really easy way for you to get this data out of the decoder package, but you could parse it and then look at the first row to get the names.

[Adrian Sieber]

but you could parse it and then look at the first row to get the names

That's what I'm doing now, but it's a shame that I can't use any of elm-csv's helper functions then. I only get a List (List String) and have to do the rest myself. 😅 A parseFields as described here, would make things quite easier.

I mean, I've already written in now: https://csvcheck.surge.sh
…but maybe for the next poor soul ;-)