Using stringByReplacingOccurrencesOfString to take into consideration "words"

Question

My goal is to use stringByReplacingOccurrencesOfString to replace occurrences of words or phrases with replacements. The words and their replacements are found in a dictionary such as that the word or phrases are keys, and their values are their replacements:

{"is fun" : "foo",
 "funny" : "bar"}

Because stringByReplacingOccurrencesOfString is literal and disregards "words" in the convention Western language sense, I am running in the trouble where the following sentence:

"He is funny and is fun",

the phrase "is fun" is actually detected twice using this method: first as part of "is funny", and the second as part of "is fun", causing an issue where a literal occurrence is used for word replacement, and not realizing that it is actually part of another word.

I was wondering if there is a way to use stringByReplacingOccurrencesOfString that takes into consideration of wording, and so a phrase like "is funny" can be viewed in its complete self, and not also be viewed as "is funny" where "is fun" detected.

By the way, this is the code I am using for replacement when iterating across all the keys in the dictionary:

NSString *newText = [wholeSentence stringByReplacingOccurrencesOfString:wordKey withString:wordValue options:NSLiteralSearch range:[wholeSentence rangeOfString:stringByReplacingOccurrencesOfString:wordKey]];
        iteratedTranslatedText = newText;

Edit 1: Using the suggested solutions, this is what I have done:

NSString *string = @"Harry is fun. Shilp is his fun pet dog";
NSRegularExpression *regex = [NSRegularExpression regularExpressionWithPattern:@"\bis fun\b" options:0 error:nil];
if (regex != nil) {
    NSTextCheckingResult *firstMatch = [regex firstMatchInString:string options:0 range:NSMakeRange(0, string.length)];
    //firstMatch is returning null
    if (firstMatch) {
        NSRange resultRange = [firstMatch rangeAtIndex:0];
        NSLog(@"first match at index:%lu", (unsigned long)resultRange.location);

    }
}

However, this is returning firstMatch as null. According to the regex tutorial on word boundaries, this is how to anchor a word or phrase, so I am unsure why its not returning anything. Help is appreciated!

For this you need to venture into the territory of *regular expressions* (`NSRegularExpression`) — borrrden, Oct 15 '14 at 03:46
Thanks for the tip @borrrden. Would you say that the answer recommended here would be the suggest way to go about it? http://stackoverflow.com/questions/9661690/user-regular-expression-to-find-replace-substring-in-nsstring — daspianist, Oct 15 '14 at 03:50
You would do better to use `NSScanner` for this, especially if you have a lot of replacements. Looping over the string once for each replaceable will be time-consuming. With a scanner, you only go through the string once. See e.g. http://stackoverflow.com/a/21100435/ or https://github.com/woolsweater/NSString-WSSHTMLEntityConversion (compare the performance of that code with the repo linked in the header to see what I'm talking about). — jscs, Oct 15 '14 at 03:50
@JoshCaswell this is great advice about `NSScanner`, and its something that I haven't not heard about before. Thanks for pointing it out! — daspianist, Oct 15 '14 at 03:59
Oh, actually, looking at that GitHub code for the first time in quite a while, I think the major performance benefit may have been using `NSMutableString` instead of `NSString`. But it does have some `NSScanner` code for demonstration. — jscs, Oct 15 '14 at 04:05

score 1 · Accepted Answer · answered Oct 15 '14 at 04:10

1

As your comment, you can use NSRegrlarEXPression in your project. For example:

NSString *string = @"He is funny and is fun";
NSRegularExpression *regex = [NSRegularExpression regularExpressionWithPattern:@"is fun([^a-zA-Z]+|$)" options:0 error:nil];
if (regex != nil) {
    NSTextCheckingResult *firstMatch = [regex firstMatchInString:string options:0 range:NSMakeRange(0, string.length)];
    if (firstMatch) {
        NSRange resultRange = [firstMatch rangeAtIndex:0];
        NSLog(@"first match at index:%d", resultRange.location);
    }
}

And to result: first match at index:16

answered Oct 15 '14 at 04:10

Pandara

139
7

1

This would also match "his fun pet dog" etc. A better regex would probably be `/bis fun/b` – borrrden Oct 15 '14 at 05:59
Thanks for chiming in. @borrrden I tried the suggestion of just using `/bis fun/b` as the pattern but am not getting any range matches using the code Pandara provided. Should I be using `/bis fun/b([^a-zA-Z]+|$)`? – daspianist Oct 15 '14 at 16:58
Thanks for answering @Pandara. This solution is very close - per the comment, the `is fun([^a-zA-Z]+|$)` doesn't work for sentences like "his fun pet", but I tried `/bis fun/b` `firstMatch` is returning nil. What should I modify the regex to improve the accuracy? Thanks! – daspianist Oct 15 '14 at 17:10
FYI I realized that the "anchor" word is `\b` according to http://www.regular-expressions.info/wordboundaries.html but unfortunately it still did not work. – daspianist Oct 15 '14 at 18:31
2

Oh man, I typed forward slash instead of backslash. Be aware that you need to escape it when you use it in Xcode (i.e. type `\\b` instead of `\b`) – borrrden Oct 16 '14 at 00:44

Using stringByReplacingOccurrencesOfString to take into consideration "words"

1 Answers1