How can I split a command string using regex

Question

I have to split a command string into segments using regex. I am looking for a very basic parser to create some custom functions, for example I have this command:

rm --remove all --keep some --but-not *.php --or-like "?-imp-*.*"

Now I want to split this string into multiple segments each containing the argument name and value, e.g.

rm
--remove all
--keep some
--but-not *.php
--or-like "?-imp-*.*"

So I can further split each segment off blank space and have the argument name and value separated.

I am not good at RegEx. So far I've written this Regex to extract the argument and value part only but it does not match the words at end of string or those with special characters like * and ?

Regex

(?<=\s)--([^--]*)(?=(\s--))

and then I grab the name of command by

(^\w+)

Any thought on this ?

Does it all really have to be done with a single regex? You could just split from spaces, loop through the result and combine anything that starts with -- with the next element (if the next element doesn't start with --). — JJJ
– JJJ, Commented Jan 26, 2013 at 11:24
I don't think this is a job for (a single) regex. There should be existing library for this. — nhahtdh
– nhahtdh, Commented Jan 26, 2013 at 11:37
@beny23 Does language matters for Regex ? I want to use it in javascript and in a shell script as well. — Gufran
– Gufran, Commented Jan 26, 2013 at 11:53
@Juhana splitting into array and joining the elements does not seems reliable to me. It might fail in some cases. — Gufran
– Gufran, Commented Jan 26, 2013 at 11:54

Billy Moon · Accepted Answer · 2013-01-26 12:58:08Z

1

It is not a good idea to use regex for parsing, but regex should be used for tokenising.

Having said that, here is an imperfect regex that matches your scenario (but not all use cases)

Implemented in javascript...

str = 'rm --remove all --keep some --but-not *.php --or-like "?-imp-*.*"'   
regex = /(^\w+\b|--[\w-]+(\s([\w*.]+|".+?"))?)/g
res = str.match(regex)
// ['rm','--remove all','--keep some','--but-not *.php','--or-like "?-imp-*.*"']

Each item will need further processing to split into keys and values.

answered Jan 26, 2013 at 12:58

Billy Moon

58.9k27 gold badges148 silver badges244 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Gufran Over a year ago

I feel jealous of you guys :( Thanks for the help, I find this one much better than other solutions.

nhahtdh Over a year ago

@Gufran: Since the answer doesn't state the assumption: command is assumed to only contain A-Z,a-z,0-9,_ (command is allowed to be something like 234__9). The name of the switch is assumed to only contain A-Z,a-z,0-9,_ and - (allowed to be this ---). The value can either be without quote or with quote. If without quote, it can only contain A-Z,a-z,0-9,_, and * and .. If with quote, it can contain any character besides new line, but the problem is that you won't be able to specify " as argument.

MikeM · Accepted Answer · 2013-01-26 15:27:53Z

1

Example implementation in Javascript:

var match,
    str = 'rm --remove all --keep some --but-not *.php --or-like "?-imp\'\'-*.*"',
    args = [],
    reg = /\s--(\S+)\s+((["']).*?[^\\]\3|\S+)/g;

while ( match = reg.exec( str ) ) {
    args.push( [ match[1], match[2] ] );
}

console.log( args );

/* 
    [ [ "remove", "all" ], [ "keep", "some" ],
      [ "but-not", "*.php" ], [ "or-like", ""?-imp-*.*"" ] ]
 */

Note: this is not intended to be fully watertight and it requires that the format of a command is validated first.

Known limitation: an argument's value may not start with a quote mark and not have a closing quote mark, e.g. vales such as " and 'n will break the parse.

edited Jan 26, 2013 at 15:27

answered Jan 26, 2013 at 12:47

MikeM

13.8k3 gold badges38 silver badges49 bronze badges

8 Comments

nhahtdh Over a year ago

[^\3] --> You cannot do this. It will match any character, except for the character with ASCII code 3.

nhahtdh Over a year ago

Same problem as Billy Moon's solution regarding the quoted text, though.

MikeM Over a year ago

@nhahtdh. Added known limitation. Further criticisms appreciated.

nhahtdh Over a year ago

You can have a look at this (don't copy over, though, since the OP's requirement can be different): stackoverflow.com/questions/13799773/…

MikeM Over a year ago

@nhahtdh.

"'ab\'c'".match( /'(?:[^'\\\n\r\u2028\u2029]|\\(?:[^\n\rxu0-9]|0(?![0-9])|x[0-9a-fA-F]{2}|u[0-9a-fA-F]{4}|\n|\r\n?))*'/ ); // "'ab'"

Not sure how that helps here, or how it is better than what I have used above for this use case.

|

beny23 · Accepted Answer · 2013-01-26 16:12:36Z

0

I would use library which implements GetOpt for javascript for that purpose (otherwise you're reinventing the wheel):

A quick google search brought up the following:

Note, I have not tried any of these.

answered Jan 26, 2013 at 16:12

beny23

35.1k6 gold badges84 silver badges86 bronze badges

Collectives™ on Stack Overflow

How can I split a command string using regex

3 Answers 3

2 Comments

8 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

8 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related