Count the number of occurences of substring in a string in Java

[DBQ]

Does anybody know the simplest and cleanest way of of counting the number of times a substring occurs in a string in Java?

Is there also a way to accomplishing the task using the regular expressions?

[slavik]

I believe that you can, one of the objects returns an array of matches. I don't remember exactly, but it is doable (unless I am not remembering something).

[CptPicard]

Regexps are probably the easiest one... you just call "find" as many times as it returns something, and count those.

[ghostdog74]

see here for some suggestions

[themusicwave]

You could use String.indexOf(string,start)

Like so:


Where str is the string you are searching in and findStr is what you are looking for.

lastIndex = 0;
count =0;

while(lastIndex != -1){

       lastIndex = str.indexOf(findStr,lastIndex);

       if( lastIndex != -1){
             count ++;
      }
}

That would do exactly what you need. There may be more efficient solutions out there like regular expressions though. This one is decent though.

[DBQ]

Thank you all. Everything was very helpful. I stuck with the regular expression option.

[Balazs_noob]

http://java.sun.com/docs/books/tutor...tcherDemo.java

[sav991]

       lastIndex = str.indexOf(findStr,lastIndex);

You would have to change this line to ensure subsequent iterations don't just keep looking at the previous result, but yeah, regular expressions sound like the go since I'm wanting to perform validation tasks.

Thanks.

Update:
Wanting to count the number of commas in a line of CSV text.

import java.util.regex.Pattern;

public static Boolean isValid(final String s) {
        String regex = "[^,]*,[^,]*,[^,]*,[^,]*,[^,]*";

        return Pattern.matches(regex, s);
}

[myrtle1908]

Wanting to count the number of commas in a line of CSV text.

I always liked split for this task ...

String s = "a,b,c,d";
System.out.println(s.split(",").length-1);

Sure we are creating an array but I don't care

Obviously this doesnt cater for quoted commas etc but you can mod the regex to suit.

[Some Penguin]

Update:
Wanting to count the number of commas in a line of CSV text.

Regexes might be overkill for counting commas.

If you don't care about commas in quoted strings, or any escape characters, you can just loop over the results of stringVariable.toCharArray() incrementing a variable whenever you see a comma. Linear time, constant space cost.


If you do care, something like

boolean escapeMode = false;  // \ = escape next character
boolean quoteMode  = false;  // "foo"
int count = 0;

// note:  below does necessarily not handle all 
// multi-byte Unicode characters; it also does not handle
// quote characters except "
//
// if that is a concern, use String's getCodePointAt() 
// and Character's toChars(), as well as Character's 
// getType() for identifying initial/final quotes... 

for (char c : stringVariable.toCharArray()) {
  if (escapeMode) {
     // Only good for one character.
     escapeMode = false;
  } else if (c == '\\') {
     // next character is treated as normal even if special
     // even if we're in a quote; this allows a quote within
     // a quoted string
     escapeMode = true;
  } else if (c  == '"') {
     quoteMode = !quoteMode;
  } else if ((!quoteMode) && (c == ',')) {
     count++;
  }
}