DataFrame: LoadCsv and WriteCsv

[Happypig375]

Like why not ReadCsv-WriteCsv or LoadCsv-SaveCsv?
Why no consistency here?
Even the docs say "read" for LoadCsv, but if "read" implies reading from a string, then WriteCsv does not make sense.

(But the current state is already better than Pandas' read_csv-to_csv nonsense.)

[michaelgsharp]

Is this for DataFrames? or which part are you talking about?

[Happypig375]

Deprecate either LoadCsv or WriteCsv and replace with ReadCsv or SaveCsv for consistency.

[ericstj]

@eerhardt do you recall why DataFrame went with Load/Write instead of Load/Save or Read/Write?

Looks to me like it was initially Read, then changed to Load dotnet/corefxlab#2760 then a @dcostea added Write dotnet/corefxlab#2947.

[eerhardt]

Looks to me like it was initially Read, then changed to Load

We explicitly went with Load instead of Read because you are "loading" a CSV stream into a DataFrame. Just like XDocument.Load(Stream).

The name came from the API review. See https://github.com/dotnet/apireviews/tree/main/2019/11-05-dataframe.

If the opposite word of Load is Save, I think we can rename WriteCsv to SaveCsv. I don't think the write API ever went through an API review, since it happened afterwards.

[rosebyte]

I'm sure we won't bring breaking changes because of this.