Better CSV support #53
Conversation
This was added to support converting data loaded from a CSV as Strings to numeric data. However, the new CSV parser just creates 2 columns, one for numeric data and one for textual data, then `orElse`s them together.
Conflicts: framian/src/main/scala/framian/Frame.scala framian/src/main/scala/framian/csv/CsvRowExtractor.scala
Conflicts: framian/src/main/scala/framian/FrameUtils.scala framian/src/main/scala/framian/UntypedColumn.scala framian/src/main/scala/framian/csv/CsvRowExtractor.scala
| /** | ||
| * A single row in a CSV file. | ||
| */ | ||
| final class CsvRow(val cells: Vector[CsvCell]) extends AnyVal { |
There was a problem hiding this comment.
This is mostly used in a Vector of Either, so I guess it will end up being boxed most of the time. Do you think we could make it a type alias instead?
There was a problem hiding this comment.
Well, I'd still prefer a value class over a type alias. Makes implicits and whatnot easier and is a much more real. Is there any reason to make it a type alias?
There was a problem hiding this comment.
It doesn't matter much to me, I just thought that you wanted to avoid boxing as much as possible, and this AnyVal will be end up being boxed most of the time, based on what I see in the rest of the code.
There was a problem hiding this comment.
Oh, yes, I see :) Argh... yeah, I'm not sure the best solution. I'm not terribly worried about boxing here, since so much boxing already goes into making a row that I don't think it would make much of a difference. I'll ponder, but perhaps as a separate issue.
|
Looks very good to me. Feel free to merge when you have figured the merge out :-) |
This adds a new CSV loader/exporter for Framian. This supports a feature rich set of formatting options - including row-delimiters, empty/invalid markers, and quote escapes. It also allows auto-detection of all or some of the formatting options (excluding header detection, which was flakey). The parser itself parses in chunks and can read the data from a stream of bytes, rather than all-at-once.