OpenCSV reader strips ending quotes instead of ignoring them

Question

Say I have:

id,description,amount
1,Foo "bar",10.5
2,Quick "brown" fox,9.0

I know that proper csv should have Foo "bar" escaped as Foo ""bar"" to pick up the quotes. But this is the data I have to deal with -- 🤷, and it is not something I can modify before processing.

try (CSVReader csvReader = new CSVReaderBuilder(new FileReader(resourcePath))
        .withSkipLines(1)
        .withCSVParser(new CSVParserBuilder().withIgnoreQuotations(true).build())
        .build()) {
    String[] line;
    ..

To solve this I try to make CSVReader to ignore all quotes with withIgnoreQuotations(true), but it seems to strip the last quotation instead of ignoring it, so the output ends up with

1,Foo "bar,10.5

. Is there any way to achieve below with OpenCSV?

1,Foo "bar",10.5

Answer 1

You can use something like the following:

new CSVParserBuilder().withQuoteChar('§').build();

Obviously this is not ideal, as you have to choose a character which is guaranteed never to appear in your data. I chose the section symbol §- that may not work for you.

Just out of interest, the Apache Commons CSV parser does not exhibit this behavior:

<dependency>
    <groupId>org.apache.commons</groupId>
    <artifactId>commons-csv</artifactId>
    <version>1.8</version>
</dependency>

And:

String sampleRecord = "1,Foo \"bar\",10.5";
CSVParser parser = CSVParser.parse(sampleRecord, CSVFormat.DEFAULT);
        
for (CSVRecord record : parser) {
    System.out.println(record.get(1));
}

This prints:

Foo "bar"