(PHP) Parse command

Question

I want to get values of command tags (GET, FROM, IN, etc.) My command is:

// My command
$_cmd = 'GET a, b FROM p IN a and c="I am from Sarajevo" or d>1 ';

// My parser
if(preg_match_all('/(GET|FROM|IN)\s+([^\s]+)/si',$_cmd, $m))
    $cmd = array_combine($m[1], $m[2]);

Output:

Array
(
  [GET] => a,
  [FROM] => p
  [IN] => a
  [from] => Sarajevo"
)

I am looking for this output:

Array
(
  [GET] => a, b
  [FROM] => p
  [IN] => a and c="I am from Sarajevo" or d>1
)

As you see, problem is with whitespaces and repeated command tags in strings (like from). So how can I parse this command?

What if you skip the /i-modifier - wouldn't you be able to make sure that you're only counting FROM and not from? Would there be a scenario where the IN would involve a string with a uppercased from? — karllindmark
– karllindmark, Commented Aug 12, 2011 at 13:26
I guess that relates to SQL... simply use your SQL, don't try to develop an interface for that. — Dor
– Dor, Commented Aug 12, 2011 at 13:29
@Dor - Please don't advice SQL parser. I need to parse this for my ORM project. — kuzey beytar
– kuzey beytar, Commented Aug 12, 2011 at 13:30
@dino beytar: "Object-relational mapping (ORM) systems (and the “frameworks” that use them) are another frequent performance nightmare.", From the book High Performance MySQL, Second Edition — Dor
– Dor, Commented Aug 12, 2011 at 13:59
@Dor - Yes, you right. But I think, this command line helps me to get standardisation my project. So i suppose, i will get less trouble with upgrades. I just want to try this way. Maybe I am completely on wrong way. — kuzey beytar
– kuzey beytar, Commented Aug 12, 2011 at 14:07

mario · Accepted Answer · 2011-08-12 13:29:58Z

8

You cannot easily parse that with a single regex. (It's doable, but not simple.)

You should use a simple tokenizer, where a regex again becomes a useful tool:

  preg_match_all('/\w+|".*?"|\W/', $_cmd = 'GET a, b FROM p IN a and c="I am from Sarajevo" or d>1 ', $list);

This gives you a simple list, where you just have to find the clauses that you are interested in, then remerge the subsequent tokens (though I'm confused about your use case):

[0] => Array
    (
        [0] => GET
        [1] => a
        [2] => ,
        [3] => b
        [4] => FROM
        [5] => p
        [6] => IN
        [7] => a
        [8] => and
        [9] => c
        [10] => =
        [11] => "I am from Sarajevo"
        [12] => or
        [13] => d
        [14] => >
        [15] => 1
    )

answered Aug 12, 2011 at 13:29

mario

146k20 gold badges243 silver badges293 bronze badges

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

4 Comments

kuzey beytar Over a year ago

I got empty spaces after each item. Can you check your code one more, please? Am i wrong?

mario Over a year ago

You can use (?!\s)\W in place of \W to remove spaces. If you want to merge the resulting parts you should keep the spaces however. (That's why I disabled that option again.) -- Really depends on if you need it as tokenizer, or just to break up the string parts.

kuzey beytar Over a year ago

Could I do this without using regex?

mario Over a year ago

Of course; but that's just more work. (String functions and lots of PHP code is typically slower than preg_match.)

SergeS · Accepted Answer · 2011-08-12 13:29:28Z

3

if( preg_match_all('/(GET|FROM|IN)(.(?!(GET|FROM|IN)))+\s*/si',$_cmd, $m))

this means - find any char after keyword which is not followed by GET, FROM or IN whith whitespace after it

answered Aug 12, 2011 at 13:29

SergeS

11.9k3 gold badges31 silver badges35 bronze badges

2 Comments

powtac Over a year ago

But GET is followed by FROM?! I see the idea, but does it work?

SergeS Over a year ago

GET ... FROM - there is intersection

Dor · Accepted Answer · 2011-08-12 13:26:15Z

1

You need to develop a scripting language for this. Regexps aren't suitable for these purposes.

answered Aug 12, 2011 at 13:26

Dor

7,5225 gold badges36 silver badges47 bronze badges

4 Comments

kuzey beytar Over a year ago

So what's your solution without using Regex?

Dor Over a year ago

@dino beytar: You could be using Regex but that's absolutely not the main part of the solution. You'll need to develop (or use an available FOSS) an interpreter: en.wikipedia.org/wiki/Interpreter_%28computing%29

kuzey beytar Over a year ago

You meant that, should i develop or use an interpreter without using PHP, or I've been already trying to develop an interpreter?

Dor Over a year ago

@dino beytar: I think that you already tried to develop one using Regex and I told you that this is not the way to develop an interpreter. But I also think that you should avoid trying to develop or use an interpreter because that will be a performance nightmare.

powtac · Accepted Answer · 2011-08-12 13:27:10Z

1

You could remove the case insensitive i after the delimiter /. And also make sure there is at least one whitespace after the keywords.

answered Aug 12, 2011 at 13:27

powtac

41.1k28 gold badges120 silver badges173 bronze badges

1 Comment

Dor Over a year ago

What if 'GET a, b FROM p IN a and c="I am FROM Sarajevo" or d>1 ' ?

Saic Siquot · Accepted Answer · 2011-08-12 13:59:44Z

1

$_cmd = 'GET a, b FROM p IN a and c="I am from Sarajevo" or d>1 ';
$tpar = preg_split('/\s+(GET|FROM|IN)\s+/i', ' '.$_cmd.' ', -1, PREG_SPLIT_DELIM_CAPTURE);
array_walk($tpar, 'trim');

print_r($tpar);

// gives:
array(
  [0] => GET
  [1] => a, b
  [2] => FROM
  [3] => p
  [4] => IN
  [5] => a and c="I am from Sarajevo" or d>1
)
// the rest is straight forward

edited Aug 12, 2011 at 13:59

answered Aug 12, 2011 at 13:49

Saic Siquot

6,5135 gold badges37 silver badges56 bronze badges

2 Comments

kuzey beytar Over a year ago

Output are not the same with you mention above. I think somethings wrong with your code? Am i wrong?

Saic Siquot Over a year ago

yes, I missed the "form" inside the quotes, I am anwsering without PHP box.

Collectives™ on Stack Overflow

(PHP) Parse command

5 Answers 5

4 Comments

2 Comments

4 Comments

1 Comment

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

5 Answers 5

4 Comments

2 Comments

4 Comments

1 Comment

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related