Backbone.js - 根据包含多个关键字的Array过滤集合

Question

I'm using Backbone.js/Underscore.js to render a HTML table which filters as you type into a textbox. In this case it's a basic telephone directory.
The content for the table comes from a Collection populated by a JSON file. A basic example of the JSON file is below:

[{
    "Name":"Sales and Services",
    "Department":"Small Business",
    "Extension":"45446",
},
{
    "Name":"Technical Support",
    "Department":"Small Business",
    "Extension":"18800",
},
{
    "Name":"Research and Development",
    "Department":"Mid Market",
    "Extension":"75752",
}]

I convert the text box value to lower case and then pass it's value along with the Collection to this function, I then assign the returned value to a new Collection and use that to re-render the page.

filterTable = function(collection, filterValue) {
    var filteredCollection;
    if (filterValue === "") {
        return collection.toJSON();
    }
    return filteredCollection = collection.filter(function(data) {
        return _.some(_.values(data.toJSON()), function(value) {
            value = (!isNaN(value) ? value.toString() : value.toLowerCase());
            return value.indexOf(filterValue) >= 0;
        });
    });
};

The trouble is that the function is literal. To find the "Sales and Services" department from my example I'd have to type exactly that, or maybe just "Sales" or "Services". I couldn't type "sal serv" and still find it which is what I want to be able to do.

I've already written some javascript that seems pretty reliable at dividing up the text into an array of Words (now updated to code in use).

toWords = function(text) {
    text = text.toLowerCase();
    text = text.replace(/[^A-Za-z_0-9@.]/g, ' ');
    text = text.replace(/[\s]+/g, ' ').replace(/\s\s*$/, '');
    text = text.split(new RegExp("\\s+"));
    var newsplit = [];
    for (var index in text) {
        if (text[index]) {
            newsplit.push(text[index]);
        };
    };
    text = newsplit;
    return text;
};

I want to loop through each word in the "split" array and check to see if each word exists in one of the key/values. As long as all words exist then it would pass the truth iterator and get added to the Collection and rendered in the table.

So in my example if I typed "sal serv" it would find that both of those strings exist within the Name of the first item and it would be returned.
However if I typed "sales business" this would not be returned as although both the values do appear in that item, the same two words do not exist in the Name section.

I'm just not sure how to write this in Backbone/Underscore, or even if this is the best way to do it. I looked at the documentation and wasn't sure what function would be easiest.

I hope this makes sense. I'm a little new to Javascript and I realise I've dived into the deep-end but learning is the fun part ;-)
I can provide more code or maybe a JSFiddle if needed.

Answer 1

Using underscore's any and all make this relatively easy. Here's the gist of it:

  var toWords = function(text) {
    //Do any fancy cleanup and split to words
    //I'm just doing a simple split by spaces.
    return text.toLowerCase().split(/\s+/);
  };

  var partialMatch = function(original, fragment) {   
    //get the words of each input string
    var origWords = toWords(original + ""), //force to string
        fragWords = toWords(fragment);

    //if all words in the fragment match any of the original words, 
    //returns true, otherwise false
    return _.all(fragWords, function(frag) {
      return _.any(origWords, function(orig) {
        return orig && orig.indexOf(frag) >= 0;
      });
    });
  };

  //here's your original filterTable function slightly simplified
  var filterTable = function(collection, filterValue) {
      if (filterValue === "") {
          return collection.toJSON();
      }
      return collection.filter(function(data) {
          return _.some(_.values(data.toJSON()), function(value) {
              return partialMatch(value, filterValue);
          });
      });
  };

Note: This method is computationally pretty inefficient, as it involves first looping over all the items in the collection, then all the fields of each item, then all words in that item value. In addition there are a few nested functions declared inside loops, so the memory footprint is not optimal. If you have a small set of data, that should be OK, but if needed, there's a number of optimizations that can be done. I might come back later and edit this a bit, if I have time.

/code samples not tested