Java : Splitting a String using Regex

Question

I have to split a string using comma(,) as a separator and ignore any comma that is inside quotes(")

fieldSeparator : ,
fieldGrouper : "

The string to split is : "1","2",3,"4,5"

I am able to achieve it as follows :

String record = "\"1\",\"2\",3,\"4,5\"";
String[] tokens = record.split(",(?=([^\"]*\"[^\"]*\")*[^\"]*$)");

Output :

"1"
"2"
3
"4,5"

Now the challenge is that the fieldGrouper(") should not be a part of the split tokens. I am unable to figure out the regex for this.

The expected output of the split is :

I think that doing this char-by-char will actually be more readable and definitely faster. And the algorithm is as simple as it gets. And it's easier to handle the "" exception which will likely appear sooner or later. — Dariusz
– Dariusz, Commented Mar 7, 2016 at 12:00
May we ask why you are working with malformed pseudo JSON input? The funkyness with the quotes makes this hard to deal with and it might be better for you to clean up the source. — Tim Biegeleisen
– Tim Biegeleisen, Commented Mar 7, 2016 at 12:06

Enteleform · Accepted Answer · 2016-03-07 12:32:52Z

4

Update:

String[] tokens = record.split( "(,*\",*\"*)" );

Result:
Image Link

Initial Solution:
( doesn't work @ .split method )

This RexEx pattern will isolate the sections you want:
(?:\\")(.*?)(?:\\")

It uses non-capturing groups to isolate the pairs of escaped quotes, and a capturing group to isolate everything in between.

Check it out here: Live Demo

edited Mar 7, 2016 at 12:32

answered Mar 7, 2016 at 11:59

Enteleform

3,8531 gold badge18 silver badges30 bronze badges

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

4 Comments

Wiktor Stribiżew Over a year ago

This regex does not match 3 or any other values not enclosed with "...".

Enteleform Over a year ago

@WiktorStribiżew I updated the solution, but in my initial solution I assumed that the "#" pattern was consistent. I didn't realize that 3 was not captured, and still wonder if @rvd purposely has a different format for 3. Either way, the new solution works.

Mikhailov Valentin Over a year ago

Sorry but your second sollution will not work for input like 1,2 when 1 and 2 are separate numbers.

Enteleform Over a year ago

@MikhailovValentine The output matches @rvd's requirements. See: Original Post / The expected output of the split is :

Wiktor Stribiżew · Accepted Answer · 2016-03-07 12:16:14Z

2

My suggestion:

"([^"]+)"|(?<=,|^)([^,]*)

See the regex demo. It will match "..." like strings and capture into Group 1 only what is in-between the quotes, and then will match and capture into Group 2 sequences of characters other than , at the start of a string or after a comma.

Here is a Java sample code:

String s = "value1,\"1\",\"2\",3,\"4,5\",value2";
Pattern pattern = Pattern.compile("\"([^\"]+)\"|(?<=,|^)([^,]*)");
Matcher matcher = pattern.matcher(s);
List<String> res = new ArrayList<String>();
while (matcher.find()){                      // Run the matcher
    if (matcher.group(1) != null) {          // If Group 1 matched
        res.add(matcher.group(1));           // Add it to the resulting array
    } else {
        res.add(matcher.group(2));           // Add Group 2 as it got matched
    }
} 
System.out.println(res); // => [value1, 1, 2, 3, 4,5, value2]

answered Mar 7, 2016 at 12:16

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

1 Comment

Tim Biegeleisen Over a year ago

The better suggestion is that he clean up his source data IMHO.

m.cekiera · Accepted Answer · 2016-03-09 12:42:23Z

1

I would try with this kind of workaround:

String record = "\"1\",\"2\",3,\"4,5\"";
record = record.replaceAll("\"?(?<!\"\\w{1,9999}),\"?|\""," ");
String[] tokens = record.trim().split(" ");
for(String str : tokens){
    System.out.println(str);
}

Output:

answered Mar 9, 2016 at 12:42

m.cekiera

5,3935 gold badges24 silver badges35 bronze badges

1 Comment

rvd Over a year ago

I ultimately had to use similar workaround, i.e, first split and then remove quotes(if present) from each token.

Matt · Accepted Answer · 2016-03-07 12:38:07Z

0

My proposition:

record = record.replaceAll("\",", "|");
record = record.replaceAll(",\\\"", "|");
record = record.replaceAll("\"", "");

String[] tokens = record.split("\\|");

for (String token : tokens) {
   System.out.println(token);
}

edited Mar 7, 2016 at 12:38

answered Mar 7, 2016 at 12:02

Matt

1,3181 gold badge12 silver badges33 bronze badges

Collectives™ on Stack Overflow

Java : Splitting a String using Regex

4 Answers 4

4 Comments

1 Comment

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

4 Comments

1 Comment

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related