Feature request: Differentiate between empty string and null value

[doyaaaaaken]

Feature request: Differentiate between empty string and null value

[StefRe]

Thanks for your reply.
RFC 4180 just describes the file format with empty fields being explicitly allowed:

field = (escaped / non-escaped)
escaped = DQUOTE *(TEXTDATA / COMMA / CR / LF / 2DQUOTE) DQUOTE
non-escaped = *TEXTDATA

However, it doesn't say anything about how to interpret the data, specifically it doesn't mention if an empty field should be an empty string or null.

For my use case (a mock ResultSet backed by csv data) it is essential to differentiate between an empty string "" and null.

OpenCSV ("com.opencsv:opencsv:5.4") provides fine-grained configuration of what the parser should consider null:

EMPTY_SEPARATORS - two sequential separators are null.
EMPTY_QUOTES - two sequential quotes are null
BOTH - both are null
NEITHER - default. Both are considered empty string.

and allows to do exactly what I need:

import com.opencsv.CSVReaderHeaderAware
import com.opencsv.CSVReaderHeaderAwareBuilder
import com.opencsv.enums.CSVReaderNullFieldIndicator

val csv = """
    "empty col","null col"
    "",
""".trimIndent()

fun main() {
    csv.byteInputStream().bufferedReader().use { br ->
        (CSVReaderHeaderAwareBuilder(br)
            .withFieldAsNull(CSVReaderNullFieldIndicator.EMPTY_SEPARATORS)
            .build() as CSVReaderHeaderAware)
            .use { cr ->
                val csvData = cr.readMap()
                assert(csvData["empty col"] == "")
                assert(csvData["null col"] == null)
            }
    }
}

So I can switch to OpenCSV but it's a bit heavy-weight as it does far more than I actually need so I'd prefer to stick with your lib for Kotlin projects.

[lucasgust]

I really want this feature...

[doyaaaaaken]

@StefRe
Thank you for your feedback!!
I'll think about this on weekend, thanks.

[doyaaaaaken]

@StefRe
Thank you for your input.
In your case, you want to regard "" as empty string and `` as null value.

It's maybe possible to have a case like the one you encountered.
However, I looked at the RFC 4180 and considered carefully, then I've felt that your case is a special case in terms of the CSV specification.

I'm very sorry, but I'll wait to deal with it until I get same feedback from others.

[doyaaaaaken]

@StefRe
Well, thank you for the good suggestions.
I'm convinced of your opinion.
Especially that configuration of OpenCSV seems to be also applicable for kotlin-csv.
I'd like to consider this topic as a feature request and try to implement it.