Regular Expression Negate Colon doesn't work

Question

I have a JSON String which is:

{"dependencies":["xz","pkg-config","glib","gobject-introspection"],"conflicts_with":[],"caveats":null,"options":[{"option":"--universal","description":"Build a universal binary"}]}

And I wrote a regular expression to find the array behind "dependencies":

(?<=\"dependencies\":).*[^:](?=,)

in Java:

"(?<=\\\"dependencies\\\":).*[^:](?=,)"

However the result turns out:

["xz","pkg-config","glib","gobject-introspection"],"conflicts_with":[],"caveats":null,"options":[{"option":"--universal"

And only the last colon was excluded.

Please help me out.

@stribizhew yes and it turns out " ["xz" ", It seems like [^:} do not take any effectToT — spacegoing, Apr 08 '15 at 12:03
And what if you use `(?<=\\\"dependencies\\\":).*?(?=(?<=\\]),)` ? — Wiktor Stribiżew, Apr 08 '15 at 12:06
The question is " [^:] " doesn't exclude " ] ". Even that expression returns the "right" result but is not able to be generalized. I want to find the value behind a given key. Not every value is an array so...:P — spacegoing, Apr 08 '15 at 12:10
Please post the expected result. If you want to exclude `[` and `]`, you can use `(?<=\\\"dependencies\\\":\\[).*?(?=\\],)` regex. — Wiktor Stribiżew, Apr 08 '15 at 12:11
@SashaSalauyou yes this can also get the right result. But could u please explain why the negation on : doesn't work in my expression? — spacegoing, Apr 08 '15 at 12:13
@stribizhev The thing is I want to get the value for any given key. The regex I came up with is to find the longest sequences ended with , but is excluded with :. This is what [^:](?=,) supposed to mean. But why the negation on : doesn't work? Many thanks! — spacegoing, Apr 08 '15 at 12:18
@ChangLi because `.*` part takes as much characters as possible with greedy search turned on by default. `.*[^:]` means "any number of any symbols where not a colon is at the end", for example `abc:bca` satisfies it. — Alex Salauyou, Apr 08 '15 at 12:18
I guess you wanted `[^:]*` to match any character but `:` any number of times. `.*` is greedy. — Wiktor Stribiżew, Apr 08 '15 at 12:20

score 2 · Accepted Answer · answered Apr 08 '15 at 12:22

2

I suggest using this regex:

(?<=\"dependencies\":).[^:]*(?=,)

Or, almost equal:

(?<=\"dependencies\":)[^:]+(?=,)

answered Apr 08 '15 at 12:22

Wiktor Stribiżew

607,720
39
448
563

However, I think the `.` is unnecessary, it is already matched by `[^:]*`. `(?<=\"dependencies\":)[^:]+(?=,)` is enough. `+` will require at least 1 character in the match. – Wiktor Stribiżew Apr 08 '15 at 12:24

Marchev · Answer 2 · 2015-04-08T13:03:14.167

1

You could use a non-greedy zero-or-more quantifier:

(?<=\"dependencies\":)\[(.*?)\]

This would match ["xz","pkg-config","glib","gobject-introspection"] in the provided JSON.

edited Apr 08 '15 at 13:03

answered Apr 08 '15 at 12:22

Marchev

1,340
1
10
11

score 0 · Answer 3 · edited May 23 '17 at 11:58

This brought to mind parsing HTML with a regular expression. Have you tried using a JSON Parser instead?

String jsonString = "{\"dependencies\":[\"xz\",\"pkg-config\",\"glib\",\"gobject-introspection\"],\"conflicts_with\":[],\"caveats\":null,\"options\":[{\"option\":\"--universal\",\"description\":\"Build a universal binary\"}]}";
JsonReader jsonReader = Json.createReader(new StringReader( jsonString ));
JsonObject obj = jsonReader.readObject();
jsonReader.close();
String dependencies = obj.getJsonArray( "dependencies" ).toString();

Regular Expression Negate Colon doesn't work

3 Answers3