Cells with embedded carriage returns are not surrounded by quotes

[lespea]

If you write a string that contains a carriage return I thought that it would be escaped in the CSV. As it is currently, excel cannot open the generated csv without errors.

Test code

import com.univocity.parsers.csv.Csv;
import com.univocity.parsers.csv.CsvWriter;

import java.nio.charset.StandardCharsets;
import java.nio.file.Files;
import java.nio.file.Paths;

public class CsvTest {
    public static void main(String[] args) throws Exception {
        final CsvWriter csv = new CsvWriter(
              Files.newBufferedWriter(Paths.get("blah.csv"), StandardCharsets.UTF_8),
              Csv.writeExcel()
        );

        try {
            csv.writeHeaders("Label", "Value", "Trailing");

            String[][] rows = {
                  {"Start newline", "\nabc123", "Trailing"},
                  {"Middle newline", "abc\n123", "Trailing"},
                  {"End newline", "abc123\n", "Trailing"},
                  {"Start cr", "\rabc123", "Trailing"},
                  {"Middle cr", "abc\r123", "Trailing"},
                  {"End cr", "abc123\r", "Trailing"},
                  {"Start both", "\n\rabc123", "Trailing"},
                  {"Middle both", "abc\n\r123", "Trailing"},
                  {"End both", "abc123\n\r", "Trailing"},
                  };

            for (String[] row : rows) {
                csv.writeRow(row);
            }
        } finally {
            csv.close();
        }
    }
}

Output

bat blah.csv
───────┬──────────────────────────────────────
       │ File: blah.csv
───────┼──────────────────────────────────────
   1   │ Label,Value,Trailing
   2   │ Start newline,"
   3   │ abc123",Trailing
   4   │ Middle newline,"abc
   5   │ 123",Trailing
   6   │ End newline,"abc123
   7   │ ",Trailing
   8   │ Start cr,^Mabc123,Trailing
   9   │ Middle cr,abc^M123,Trailing
  10   │ End cr,abc123,Trailing
  11   │ Start both,"
  12   │ ^Mabc123",Trailing
  13   │ Middle both,"abc
  14   │ ^M123",Trailing
  15   │ End both,"abc123
  16   │ ",Trailing
───────┴──────────────────────────────────────

[jbax]

Thanks I'll fix this asap.

By default it will only escape the \r or \n if these have been configured in the CsvFormat.

CsvWriterSettings settings = Csv.writeExcel(); //format is set with line ending = `\r\n`
settings.getFormat().getnormalizedNewline(); // defaults to `\n`

The normalizedNewLine is the key here. If found on the value being written, it will be transformed into the line separator defined in the format. A value like \nabc will be written to the output as "\r\nabc". It won't do anything to \rabc though.

An additional setting is required to make the writer quote the field if it finds this \r

settings.setQuotationTriggers('\r');

Now \rabc should be written as "\r\nabc" .

If you want to have the field contain the original line ending, and ensure the value is written as "\rabc", you need an additional setting:

settings.setNormalizeLineEndingsWithinQuotes(false);

However I don't think Excel will like that one.

Hope this helps. CSV is a mixed bag and supporting all variations of intended input/output formats makes the API a bit more complex to use.

[jbax]

Done. I've released a 2.8.2-SNAPSHOT build you can use to test this. Thank you for using our parsers!

[lespea]

2.8.2 works perfect; thanks for the quick reply!

bat blah.csv
───────┬──────────────────────────────────────
       │ File: blah.csv
───────┼──────────────────────────────────────
   1   │ Label,Value,Trailing
   2   │ Start newline,"
   3   │ abc123",Trailing
   4   │ Middle newline,"abc
   5   │ 123",Trailing
   6   │ End newline,"abc123
   7   │ ",Trailing
   8   │ Start cr,"^Mabc123",Trailing
   9   │ Middle cr,"abc^M123",Trailing
  10   │ End cr,"abc123",Trailing
  11   │ Start both,"
  12   │ ^Mabc123",Trailing
  13   │ Middle both,"abc
  14   │ ^M123",Trailing
  15   │ End both,"abc123
  16   │ ",Trailing
───────┴─────────────────────────────────────