How can I match the names in an array such as this one:
[milk, pumpkin pie, eggs, juice]
It must also support single items in arrays. This works, but is crashing Reggy (a regex program) probably because the constant lookaheads are a strain.
([^,\[\]]+(?=,|\s*\]))
Also note that I don't want to capture the commas or square brackets - just the items...
EDIT:
I've gotten: (?<=\[)([^\[\]]+)(?=,|\]) which matches the items and commas.
For all I can tell (and I tried using Python and its built-in regular expressions), there is nothing wrong with your regex. If it causes Reggy to crash, that is probably a bug and should reported as such.
However, it should be noted that your regex, while it avoids the commas and brackets to appear inside the matches, does include the spaces between a comma and the beginning of an item. For example, you will get " pumpkin pie" (note the leading space), rather than "pumpkin pie" as a match. I don't see any direct way to avoid this.
One way, but possibly not supported by Reggy, is to use groups to sub-select the relevant parts of matches. For example in Python:
import re
text = '[milk, pumpkin pie, eggs, juice]'
pattern = re.compile(r'\s*([^,\[\]]+)(?=,|\s*\])')
for match in pattern.finditer(text):
print match.group(1)
Note how the regex now includes leading whitespace (\s+) and round brackets around the relevant part of the match: ([^,\[\]]+). In the printing part I refer to this as group(1).
Here's what I would suggest in .NET:
(?<=\[(?:[^\]]+,\s+)?) // Look behind for the start bracket and possibly previous values
([^\],]+) // capture the value until the next comma or end bracket
(?=,|]) // Look ahead and find a comma or end bracket
(Broken into multiple lines for clarity only.)
The issue with using JavaScript's variant of regex is the lack of a zero-width positive look-behind assertion, which is needed if you want to match more than one element of the array.
javascript lookahead works fine, and this doesn't capture the comma/space sequences:
//only assumption is csv
var data = '[milk, pumpkin pie, eggs, juice]';
var myregexp = /[^,]+(?=,\s?|]$)/g;
var match = myregexp.exec(data);
var result = "matches found:\n";
while (match != null) {
//first match will include the [, strip it.
result += match[0] + ',\n';
match = myregexp.exec(data);
}
alert(result);
[...], thanks thoughTry \b[\w ]+\b.
This will match multiple words allowing spaces in between ([\w ]+). Since the + is greedy it will match as many words as possible, but it won't go over a comma or bracket boundary because that doesn't match \w or .
You can play around with it here.
([^,\[\]\s]+)? The space is special?[and,.