Anyone any good at regexp?

[DiscoStu]

And I don't feel like I've drip fed anything. My original question was, can someone give me some regexp to parse this specific string. No more, no less.

Is there anything you're good at except getting canned?

[OwlHoot]

Yeah I was a PM so have no technical knowledge at all...

but when will it be completed by I have a box to tick?

If you have any say in the format of the incoming string, I'd try and steer them into using a CSV format from which each field can be split by using the built-in C# CSV Parser or a 3rd party CSV parser.

You may even be able to specify the delimiter character with those. (The default will be commas, but you may prefer colons.)

Use of a proper CSV parser avoids issues with quoted delimiters, e.g. "a,b","c","d" which actually comprise only the three fields "a,b" and "c" and "d". It also even allows multi-line fields.

The string parsing, however you do it, and the database handling are two different issues. So I would try and keep those conceptually separate.

edit: I think CSV even supports two levels of delimiter. So you could parse something like "a=1,b=3,c=4" into a handy structure by specifying "," as the top-level delimiter and "=" as the next level delimiter. But that said, CSV means different things to different people and probably not all parser implementations support multi-level parsing.

[suityou01]

No you didn't

All in post #4

HTH

[suityou01]

If you have any say in the format of the incoming string, I'd try and steer them into using a CSV format from which each field can be split by using the built-in C# CSV Parser or a 3rd party CSV parser.

You may even be able to specify the delimiter character with those. (The default will be commas, but you may prefer colons.)

Use of a proper CSV parser avoids issues with quoted delimiters, e.g. "a,b","c","d" which actually comprise only the three fields "a,b" and "c" and "d". It also even allows multi-line fields.

The string parsing, however you do it, and the database handling are two different issues. So I would try and keep those conceptually separate.

Tricky one to sell as ClientCo want regex.

I do take your point though.

[FiveTimes]

Can't you just split on ':' ?

[d000hg]

I just use StackOverflow for Regex questions. Agree about them being a pig though.

Recently I've gotten into using regex-search in Visual C++... or even regex-find-and-replace. You can do pretty fancy stuff except MS use their own regex syntax (of course they do).

[Bunk]

I'm not even sure it's possible with a regex.

Thisisalargestring: And here is some data up until the next space character Thisisanotherlargestring: And here is some more data

"And here is some data up until the next space character" is full of spaces so how are you supposed to know that the last space is the one to stop at? Maybe because "Thisisanotherlargestring" is followed by a colon? But you don't want to include "Thisisanotherlargestring" in the match so you need to look-ahead. Now it's getting beyond my knowledge of regexes.

[OwlHoot]

Can't you just split on ':' ?

Exactly, even if that is just a first step to then using regexps for individual fields.

That's why I suggested using CSVs (which despite standing for "comma-separated values" can just as well use colons as field delimiters)

[suityou01]

Exactly, even if that is just a first step to then using regexps for individual fields.

That's why I suggested using CSVs (which despite standing for "comma-separated values" can just as well use colons as field delimiters)

And then clientco say oh, suity we want to retrieve this piece of data and it looks like :

*12345*

I'm not so sure on why everyone is so hung up on why I want to use regexp for parsing unstructured data.

Thanks for all your input, it is appreciated.

[NickFitz]

You still haven't said how you want to parse it. Do you want to extract all uppercase letters? All hexadecimal digits that are odd? All non-space characters?

In default of any clear specification, I shall make the assumption that FolderGUID, Outcome, and DataItem are field labels, the field value follows the label after a colon, and fields are delimited by a single space or the end of the line.

If that is the case then, in JavaScript (because I can test that in the browser, and I'm not firing up a C# compiler for a no-brainer like this) the following regular expression:

/FolderGUID:([^ ]*) Outcome:([^ ]*) DataItem:(.*$)/

will give the following array when presented with the input you specified (which is repeated as the complete match at element 0):

[
    "FolderGUID:67bfabff-ad78-4d30-918e-811dd2636f83 Outcome:Accept DataItem:SomeData",
    "67bfabff-ad78-4d30-918e-811dd2636f83",
    "Accept",
    "SomeData"
]

To generalise it into a function that takes the input as a (single line) string and returns an object with the extracted values as named properties:

function parseLineInWhatOneAssumesToBeTheRequiredWay(line) {
    var pieces = line.match(/FolderGUID:([^ ]*) Outcome:([^ ]*) DataItem:(.*$)/);
    if (pieces) {
        return {
            folderGUID: pieces[1],
            outcome: pieces[2],
            dataItem: pieces[3]
        };
    }
    return null;
}