获取字符串中某个位置周围的单词

Question

I would like to get the words that are around a certain position in a string. For example two words after and two words before.

For example consider the string:

String str = "Hello my name is John and I like to go fishing and hiking I have two sisters and one brother.";
String find = "I";

for (int index = str.indexOf("I"); index >= 0; index = str.indexOf("I", index + 1))
{
    System.out.println(index);
}

This writes out the index of where the word "I" is. But I want to be able to get a substring of the words around these positions.

I want to be able to print out "John and I like to" and "and hiking I have two".

Not only single word strings should be able to be selected. Search for "John and" will return " name is John and I like".

Is there any neat, smart way of doing this?

Answer 1

Single word:

You can achiveve that using String 's split() method . This solution is O(n) .

public static void main(String[] args) {
    String str = "Hello my name is John and I like to go fishing and "+
                         "hiking I have two sisters and one brother.";
    String find = "I";

    String[] sp = str.split(" +"); // "+" for multiple spaces
    for (int i = 2; i < sp.length; i++) {
        if (sp[i].equals(find)) {
            // have to check for ArrayIndexOutOfBoundsException
            String surr = (i-2 > 0 ? sp[i-2]+" " : "") +
                          (i-1 > 0 ? sp[i-1]+" " : "") +
                          sp[i] +
                          (i+1 < sp.length ? " "+sp[i+1] : "") +
                          (i+2 < sp.length ? " "+sp[i+2] : "");
            System.out.println(surr);
        }
    }
}

Output:

John and I like to
and hiking I have two

Multi-word:

Regex is a great and clean solution for case when find is a multi-word. Due to its nature, though, it misses the cases when the the words around also match find (see the an example of this below).

The algorithm below takes care of all cases (all solutions' space). Bear in mind that, due to the nature of the problem, this solution in the worst case is O(n*m) ^{(with n being str 's length and m being find 's length)} .

public static void main(String[] args) {
    String str = "Hello my name is John and John and I like to go...";
    String find = "John and";

    String[] sp = str.split(" +"); // "+" for multiple spaces

    String[] spMulti = find.split(" +"); // "+" for multiple spaces
    for (int i = 2; i < sp.length; i++) {
        int j = 0;
        while (j < spMulti.length && i+j < sp.length 
                                  && sp[i+j].equals(spMulti[j])) {
            j++;
        }           
        if (j == spMulti.length) { // found spMulti entirely
            StringBuilder surr = new StringBuilder();
            if (i-2 > 0){ surr.append(sp[i-2]); surr.append(" "); }
            if (i-1 > 0){ surr.append(sp[i-1]); surr.append(" "); }
            for (int k = 0; k < spMulti.length; k++) {
                if (k > 0){ surr.append(" "); }
                surr.append(sp[i+k]);
            }
            if (i+spMulti.length < sp.length) {
                surr.append(" ");
                surr.append(sp[i+spMulti.length]);
            }
            if (i+spMulti.length+1 < sp.length) {
                surr.append(" ");
                surr.append(sp[i+spMulti.length+1]);
            }
            System.out.println(surr.toString());
        }
    }
}

Output:

name is John and John and
John and John and I like

Answer 2

Here is another way I found out using Regex:

        String str = "Hello my name is John and I like to go fishing and hiking I have two    sisters and one brother.";

        String find = "I";

        Pattern pattern = Pattern.compile("([^\\s]+\\s+[^\\s]+)\\s+"+find+"\\s+([^\\s]+\\s[^\\s]+\\s+)");
        Matcher matcher = pattern.matcher(str);

        while (matcher.find())
        {
            System.out.println(matcher.group(1));
            System.out.println(matcher.group(2));
        }

Output:

John and
like to 
and hiking
have two 

Answer 3

Use String.split() to split the text into words. Then search for "I" and concatenate the words back together:

String[] parts=str.split(" ");

for (int i=0; i< parts.length; i++){
   if(parts[i].equals("I")){
     String out= parts[i-2]+" "+parts[i-1]+ " "+ parts[i]+ " "+parts[i+1] etc..
   }
}

Ofcourse you need to check if i-2 is a valid index, and using a StringBuffer would be handy performance wise, if you have a lot of data ...

Answer 4

// Convert sentence to ArrayList
String[] stringArray = sentence.split(" ");
List<String> stringList = Arrays.asList(stringArray);

// Which word should be matched?
String toMatch = "I";

// How much words before and after do you want?
int before = 2;
int after = 2;

for (int i = 0; i < stringList.size(); ++i) {
    if (toMatch.equals(stringList.get(i))) {
        int index = i;
        if (0 <= index - before && index + after <= stringList.size()) {
            StringBuilder sb = new StringBuilder();

            for (int i = index - before; i <= index + after; ++i) {
                sb.append(stringList.get(i));
                sb.append(" ");
            }
            String result = sb.toString().trim();
            //Do something with result
        }
    }
}

This extracts the two words before and after the match. Could be extended to print at most two words before and after and not exactly two words.

EDIT Damn.. way to slow and no fancy ternary operators :/

Answer 5

public static void main(String[] args) {
    String str = "Hello my name is John and I like to go fishing and hiking I have two    sisters and one brother.";
    String find = "I";
    int countWords = 3;
    List<String> strings = countWordsBeforeAndAfter(str, find, countWords);
    strings.stream().forEach(System.out::println);
}

public static List<String> countWordsBeforeAndAfter(String paragraph, String search, int countWordsBeforeAndAfter){
    List<String> searchList = new ArrayList<>();
    String str = paragraph;
    String find = search;
    int countWords = countWordsBeforeAndAfter;
    String[] sp = str.split(" +"); // "+" for multiple spaces
    for (int i = 0; i < sp.length; i++) {
        if (sp[i].equals(find)) {

            String before = "";
            for (int j = countWords; j > 0; j--) {
                if(i-j >= 0) before += sp[i-j]+" ";
            }

            String after = "";
            for (int j = 1; j <= countWords; j++) {
                if(i+j < sp.length) after += " " + sp[i+j];
            }
            String searhResult = before + find + after;
           searchList.add(searhResult);
        }
    }
    return searchList;
}