CSV parser in JAVA, double quotes in string (SuperCSV, OpenCSV)

Question

all day I've been searching how to resolve this probem and nothing... I want to write function, which convert CSV file to collection of lists (of strings). Here is this function:

public Collection<? extends List<String>> parse() throws IOException {
    Collection<List<String>> collectionOfLists = new ArrayList<List<String>>();
    CsvListReader parser = new CsvListReader(Files.newBufferedReader(pathToFile, StandardCharsets.UTF_8), CsvPreference.EXCEL_PREFERENCE);

    List<String> row;
    while( (row = parser.read()) != null)
        collectionOfLists.add(row);

    return collectionOfLists;
}

public static String toString(Collection<? extends List<String>> csv) {
    StringBuilder builder = new StringBuilder();
    for(List<String> l : csv) {
        for(String s : l)
            builder.append(s).append(',');
        if(builder.length() > 0)
            builder.setCharAt(builder.length()-1,'\n');
    }
    return builder.toString();
}

But e.g. for that input:

id, name, city, age
1,"Bob",London,12

Output for toString(parse()) is:

id, name, city, age
1,Bob,London,12

instead of the same like input:/ What can I do, that strings contain \" (quotes) ? Please help me.

errr... if you want to write your own, why the "SuperCSV, OpenCSV" in the question title? — Leo
– Leo, Commented Apr 10, 2014 at 23:13
because I was trying use these classes, but I didn't find any solution — user3521479
– user3521479, Commented Apr 10, 2014 at 23:17

James Bassett · Accepted Answer · 2014-04-11 06:23:12Z

2

It's not clear from your question whether you're asking....

1. My data contains quotes - why are they being stripped out?

In this case, I'd point you to the CSV specification as your CSV file is not properly escaped, so those quotes aren't actually part of your data.

It should be

1,""Bob"",London,12

not

1,"Bob",London,12

2. How do I apply quotes when writing (even if the data doesn't contain commas, quotes, etc)?

By default Super CSV only escapes if necessary (the field contains a comma, double quote or newline).

If you really want to enable quotes, then you can configure Super CSV with a quote mode.

For example, you could always quote the name column in your example with the following preferences:

private static final CsvPreference ALWAYS_QUOTE_NAME_COL = 
    new CsvPreference.Builder(CsvPreference.STANDARD_PREFERENCE)
    .useQuoteMode(new ColumnQuoteMode(2)).build();

Alternatively, if you want to quote everything then you can use AlwaysQuoteMode, or if you want a completely custom solution, then you can write your own QuoteMode.

edited Apr 11, 2014 at 6:23

answered Apr 11, 2014 at 6:10

James Bassett

10.1k4 gold badges40 silver badges71 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

user3521479 Over a year ago

Ok, thanks. And is there any solution such that parser leaves quotes there they was, and prints strings without quotes if they wasn't in input? In sense e.g. ,, --> ,null, (or ,,), "Bob" --> "Bob", Bob --> Bob and "" --> "" Is this possible with the help of SuperCSV or OpenCSV?

James Bassett Over a year ago

There's a getUntokenizedRow() method on the readers, but that's not really what you want. I doubt many libraries will transform from CSV to Java to CSV again with identical input/output (the data will be the same, but the containing structure may differ).

sendon1982 · Accepted Answer · 2014-04-11 00:07:12Z

1

You create your own Preference.

CsvPreference excelPreference = new CsvPreference.Builder('\'', ',', "\n").build();
CsvListReader parser = new CsvListReader(Files.newBufferedReader(pathToFile , StandardCharsets.UTF_8), excelPreference);

After that, it will output as expected. In this example, you will strip the single quote if you have that in your csv file and keep the double quote untouched.

answered Apr 11, 2014 at 0:07

sendon1982

11.5k72 silver badges49 bronze badges

Comments

Sotirios Delimanolis · Accepted Answer · 2014-04-11 00:16:01Z

1

In the CsvPreference.EXCEL_PREFERENCE you've given, the quote character is the " as described in the javadoc. The quote character is a character you use to wrap special characters that want you want to appear literally.

As such, for these preferences, the appropriate way to produce your CSV content would be

id, name, city, age
1,"""Bob""",London,12

Otherwise, the CSV parser simply thinks

"Bob"

means, literally,

Bob

since there is no other special character between the quotes. But a quote is a special character so if it appears between quotes, it will be considered, literally, as a quote.

Alternatively, provide a different CsvPreference object which has a different quote character.

Make this decision only after you are certain about what your CSV producer is sending you.

edited Apr 11, 2014 at 0:16

answered Apr 10, 2014 at 23:51

Sotirios Delimanolis

281k62 gold badges718 silver badges744 bronze badges

2 Comments

user3521479 Over a year ago

Ok, I know. I mean how to convert CSV file to collection of list such that keep quotes. If you provide a different CsvPreference e.g. new CsvPreference.Builder('\'', ',', "\n").build() than, it doesn't work: id,text 1,"It doesn't work, man"

Sotirios Delimanolis Over a year ago

@user3521479 The javadoc shows how to do that. sendon1982 provides an example as well.

Collectives™ on Stack Overflow

CSV parser in JAVA, double quotes in string (SuperCSV, OpenCSV)

3 Answers 3

2 Comments

Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related