splitting a string into a multidimensional array

Question

I have a list of strings, I want to check if the string contains a specific word, and if it does split all the words in the string and add it to an associative array.

myString = ['RT @Arsenal: Waiting for the international', 'We’re hungry for revenge @_nachomonreal on Saturday\'s match and aiming for a strong finish']
wordtoFind = ['@Arsenal'] 

I want to loop through the wordtoFind and if it is in myString, split up myString into individual words and create an object like

newWord = {@Arsenal:[{RT:1},{Waiting:1},{for:1},{the:1},{international:1}]}
for(z=0; z <wordtoFind.length; z++){
 for ( i = 0 ; i < myString.length; i++) {
 if (myString[i].indexOf(wordtoFind[z].key) > -1){
 myString[i].split(" ")
 }
 }
}

• Nina Scholz: your newWord is not a valid object.
• Vincent: Can you clarify? I don't understand what you want to create. Effectively, the objet is not valid (what are the keys or {waiting:1}, {for:1}, etc. ? And you want all the words into objects like {word:1}? Can you explain?
• PostCrafter: So you want to count each of the words in your string?
• qwerty ayyy: @NinaScholz I have edited the question, is this a valid object ?
• qwerty ayyy: @Vincent sorry I am new to javascript but what i want is the main key is wordtoFind which maps onto the words in newWord and their occurrence in myString

Answer 1

I would say something likes would work, this also counts the amount of occurrences of a word in a sentence. JavaScript does not have associative arrays like PHP for instance. They just have objects or numbered arrays:

var myString = ['RT @Arsenal: Waiting for the international', 'We’re hungry for revenge @_nachomonreal on Saturday\'s match and aiming for a strong finish'];
var wordtoFind = ['@Arsenal'];
var result = {};
for(var i = 0, l = wordtoFind.length; i < l; i++) {
 for(var ii = 0, ll = myString.length; ii < ll; ii++) {
 if(myString[ii].indexOf(wordtoFind[i]) !== -1) {
 var split = myString[ii].split(' ');
 var resultpart = {};
 for(var iii = 0, lll = split.length; iii < lll; iii++) {
 if(split[iii] !== wordtoFind[i]) {
 if(!resultpart.hasOwnProperty(split[iii])) {
 resultpart[split[iii]] = 0;
 }
 resultpart[split[iii]]++;
 }
 }
 result[wordtoFind[i]] = resultpart;
 }
 }
}
console.log(result); 
//{"@Arsenal":{"RT":1,"Waiting":1,"for":1,"the":1,"international":1}}

• qwerty ayyy: thanks for this but I got a comment saying that result was not a valid object. I have now edited the question, how would this change the answer? Thanks @PierreDuc
• Poul Kruijt: @qwertyayyy Is there a reason you want it to be an array? My answer generates a valid object as well. If you check the output from the console.log you can see what it is. It would change my answer in such a way that it makes it harder (but not impossible) to count for the amount of occurrences
• qwerty ayyy: @PierreDuuc Sorry I am new to Javascript and I just got a comment saying it was an invalid object that all. But thank you for the answer
• Poul Kruijt: @qwertyayyy Well, in that case, i am sure that my answer will suffice your needs :)
• qwerty ayyy: how do I access individual elements in the result object ? Thanks

Answer 2

This method makes use of the forEach-function and callbacks. The containsWord-function was left with a for-loop for now to reduce some callbacks, this can obviously be changed.

var myString = [
 'RT @Arsenal: Waiting for the international',
 'We’re hungry for revenge @_nachomonreal on Saturday\'s match and aiming for a strong finish',
 '@Arsenal: one two three four two four three four three four'
];
var wordtoFind = ['@Arsenal'];
// define the preprocessor that is used before the equality check
function preprocessor(word) {
 return word.replace(':', '');
}
function findOccurences(array, search, callback, preprocessor) {
 var result = {};
 var count = 0;
 // calculate the maximum iterations
 var max = search.length * array.length;
 // iterate the search strings that should be matched
 search.forEach(function(needle) {
 // iterate the array of strings that should be searched in
 array.forEach(function(haystack) {
 if (containsWord(haystack, needle, preprocessor)) {
 var words = haystack.split(' ');
 // iterate every word to count the occurences and write them to the result
 words.forEach(function(word) {
 countOccurence(result, needle, word);
 })
 }
 count++;
 // once every iteration finished, call the callback
 if (count == max) {
 callback && callback(result);
 }
 });
 });
}
function containsWord(haystack, needle, preprocessor) {
 var words = haystack.split(' ');
 for (var i = 0; i < words.length; i++) {
 var word = words[i];
 // preprocess a word before it's compared
 if (preprocessor) {
 word = preprocessor(word);
 }
 // if it matches return true
 if (word === needle) {
 return true;
 }
 }
 return false;
}
function countOccurence(result, key, word) {
 // add array to object if it doesn't exist yet
 if (!result.hasOwnProperty(key)) {
 result[key] = [];
 }
 var entry = result[key];
 // set the count to 0 if it doesn't exist yet
 if (!entry.hasOwnProperty(word)) {
 entry[word] = 0;
 }
 entry[word]++;
}
// call our function to find the occurences
findOccurences(myString, wordtoFind, function(result) {
 // do something with the result
 console.log(result);
}, preprocessor);
// output:
/*
 { '@Arsenal':
 [ RT: 1,
 '@Arsenal:': 2,
 Waiting: 1,
 for: 1,
 the: 1,
 international: 1,
 one: 1,
 two: 2,
 three: 3,
 four: 4 ] }
 */

Feel free to ask any questions, if the answer needs clarification.

I hope this fits your needs.

• Poul Kruijt: You took it to the next level ;). But a little remark though, forEach is not an asynchronous function.
• PostCrafter: @PierreDuc correct, I haven't been using this too much lately, it's blocking obviously. Also thanks for the edit to the english documentation, haven't even noticed that google led me to the german version. ;)

Answer 3

You're on the right track. You just need to store the split string into the associative array variable.

var assocArr = [];
for(z=0; z <wordtoFind.length; z++){
 for ( i = 0 ; i < myString.length; i++) {
 if (myString[i].indexOf(wordtoFind[z]) > -1){
 myString[i].split(" ").forEach(function(word){
 assocArr.push(word);
 });
 }
 }
}

• mikesp: the .key is not working in the if statement - use it like this: if (myString[i].indexOf(wordtoFind[z]) > -1){.....
• senschen: Thanks. That's what I get for copy pasting I guess :/

Answer 4

I think the key problem that stuck you is the data structure. The optimal structure should be something like this:

{
 @Arsenal:[
 {RT:1, Waiting:1, for:1, the:1, international:1},
 {xxx:1, yyy:1, zzz:3}, //for there are multiple ones in 'myString' that contain the same '@Arsenal'
 {slkj:1, sldjfl:2, lsdkjf:1} //maybe more
 ]
 someOtherWord:[
 {},
 {},
 ....
 ]
}

And the code:

var result = {};
//This function will return an object like {RT:1, Waiting:1, for:1, the:1, international:1}.
function calculateCount(string, key) {
 var wordCounts = {};
 string.split(" ").forEach(function (word) {
 if (word !== key) {
 if (wordCounts[word] === undefined) wordCounts[word] = 1;
 else wordCounts[word]++;
 }
 });
 return wordCounts;
}
//For each 'word to find' and each string that contain the 'word to find', push in that returned object {RT:1, Waiting:1, for:1, the:1, international:1}.
wordToFind.forEach(function (word) {
 var current = result[word] = [];
 myString.forEach(function (str) {
 if (str.indexOf(word) > -1) {
 current.push(
 calculateCount(str, word)
 );
 }
 }); //Missed the right parenthesis here
});

• qwerty ayyy: Hi I am getting a syntax error from the statement "continue" @Yan-Yang
• Yan Yang: @qwertyayyy I'm sorry. I made a mistake. I modified the code and it's ok now. continue can't be used in forEach. Here is a question about that: stackoverflow.com/questions/31016841/…
• qwerty ayyy: Im sorry but how do I output this to the console. It just returning an empty list. @Yan-Yang
• Yan Yang: I missed a right parenthesis just now. I have added it. And I used wordToFind not wordtoFind. Also, to output to the console, use console.dir(result);.
• Yan Yang: I have tested in my console using your example of myString and wordtoFind. The result is @Arsenal: Array[1] [ 0: Object {@Arsenal:: 1, RT: 1, Waiting: 1, for: 1, international: 1, the: 1}]