capturing a string before and after some data using regular expressions in ObjectiveC

Question

I am relatively new to regex expressions and needed some advice.

The goal is to the get data in the following format into an array:

• value=777
• value=888

From this data: "value=!@#777!@#value=@#$888*"

Here is my code (Objective C):

NSString *aTestString = @"value=!@#777!@#value=@#$**888***";
NSRegularExpression *regex = [NSRegularExpression regularExpressionWithPattern:@"value=(?=[^\d])(\d)" options:0 error:&anError];

So my questions are:

1) Can the regex engine capture data that is split like that? Retrieving the "value=" removing the garbage data in the middle, and then grouping it with its number "777" etc?

2) If this can be done, then is my regex expression valid? value=(?=[^\d])(\d)

Answer 1

The lookahead (?=) is wrong here, you haven't correctly escaped the \d (it becomes \\d) and last but not least you left out the quantifiers * (0 or more times) and + (1 or more times):

NSString *aTestString = @"value=!@#777!@#value=@#$**888***";
NSRegularExpression *regex = [NSRegularExpression
    regularExpressionWithPattern:@"value=[^\\d]*(\\d+)"
    options:0
    error:NULL
];

[regex 
    enumerateMatchesInString:aTestString
    options:0
    range:NSMakeRange(0, [aTestString length])
    usingBlock:^(NSTextCheckingResult *result, NSMatchingFlags flags, BOOL *stop) {
        NSLog(@"Value: %@", [aTestString substringWithRange:[result rangeAtIndex:1]]);
    }
];

Edit: Here's a more refined pattern. It catches a word before =, then discards non-digits and catches digits afterwards.

NSString *aTestString = @"foo=!@#777!@#bar=@#$**888***";
NSRegularExpression *regex = [NSRegularExpression regularExpressionWithPattern:@"(\\w+)=[^\\d]*(\\d+)" options:0 error:NULL];

[regex 
    enumerateMatchesInString:aTestString
    options:0
    range:NSMakeRange(0, [aTestString length])
    usingBlock:^(NSTextCheckingResult *result, NSMatchingFlags flags, BOOL *stop) {
        NSLog(
            @"Found: %@=%@",
            [aTestString substringWithRange:[result rangeAtIndex:1]],
            [aTestString substringWithRange:[result rangeAtIndex:2]]
        );
    }
];

// Output:
// Found: foo=777
// Found: bar=888

Answer 2

Regular expresssions are expressions that match a given pattern. A regular expression could match, say, a string like "value=!@#777" using an expression like "value=[#@!%^&][0-9]", which says to match the literal "value=", and then any string made up of the characters #, @, !, %, ^, and &, and finally any string made up of digits. But you can't use a single regular expression by itself to get just the parts of the string that you want, i.e. "value=777".

So, one solution would be to create an expression that recognizes strings such as "value=!@#777", and then do some further processing on that string to remove the offending characters.

I think you'll be better off using NSScanner to scan the data and extract the parts you want. For example, you can use -scanString:intoString: to get the "value=" part, followed by -scanCharactersFromSet:intoString: to remove the part you don't want, and then call that method again to get the collection of digits.