Parse a string with PHP

[Sneo]

Hey,

Just a quick question on something I can't figure out to do with PHP!

 

Case 1:
$text = "You worked at the local bar and made $1456 for the night";

Case 2:
$text2 = "You worked at the local shop and made $5831 for the day";

 

Im looking for the following result:

Case 1:
$worked = "bar";
$wage = "$1456";

Case 2:
$worked = "shop";
$wage = "$5831";

 

Could someone help me out here?

[effigy]

<pre>
<?php

$text = array(
	"You worked at the local bar and made $1456 for the night",
	"You worked at the local shop and made $5831 for the day"
);

foreach ($text as $piece) {
	preg_match('/(?<=local)\s+(\S+)[^$]+(\$\d+)/', $piece, $matches);
	array_shift($matches);
	print_r($matches);
}

?>
</pre>

[Sneo]

Mmm.. Looks good! I have to change the variables to an array? It would be easier for me to keep them as variables..

[The Little Guy]

Give this a try:

<?php
$text = "You worked at the local bar and made $1456 for the night";

if(preg_match("/local.+and/",$text,$matches)){
echo $matches[0];
}else{
echo 'nothing found';
}
?>

[Sneo]

That works perfectly! Now how can i get the money amount out?

[chocopi]

just use The Little Guy's code and change it slightly

 

<?php
$text = "You worked at the local bar and made $1456 for the night";

if(preg_match("/made.+for/",$text,$matches)){
echo $matches[0];
}else{
echo 'nothing found';
}
?>

 

I think that should work

 

~ Chocopi

[Sneo]

That leaves me with:

made $1456 for

 

Anyway to get rid of the made and for now? [sorry for all these questions, first time i've had to do this with PHP]

[The Little Guy]

there may be a better way, but this works for me:

 

<?php
$text = "You worked at the local bar and made $1456 for the night";

if(preg_match("/local.+for/",$text,$matches)){
list($unsed,$worked,$unused,$unsed,$wage) = explode(" ",$matches[0]);
echo $worked.'<br>'.$wage;
}else{
echo 'nothing found';
}
?>

 

this will not work, if "bar" is changed to something containing 2+ words, like "super market"

[The Little Guy]

actually it will work with 2+ words, you will have to modify the word before you parse it, to a single word, and maybe replace the spaces with a symbol, like an underscore, then read in the word, then replace the underscores back into spaces to print out the word.

[Wildbug]

Mmm.. Looks good! I have to change the variables to an array? It would be easier for me to keep them as variables..

 

No, you don't have to change them to an array; effigy just did that in order to loop through the two options.  The important part of the code is the preg_match bit.

 

preg_match('/worked at the local (.*?) and made (\$\d+)/i', $text, $matched);
echo $matched[1];
echo $matched[2];

 

The regular expression can be as general or as specific as you need.

[Sneo]

Not a fan of that, as it isn't just based around this one text string.. There is a few, basically I'm going to be able to Post these "Earns".

 

The real example would be:

[Note: Only one would be posted at a time]

Case 1:

You walked the beat for ages today - protecting the citizens of Paris and earned yourself a $91 bonus! No taxes were taken as Paris has no mayor.

 

Case 2:

You scammed some poor suckers into your scam - they invested their last $2714 and bought some of your phoney stocks!

 

I need the script to figure out which Earn I have done. Bearing in mind, the last part: "No taxes were taken as Paris has no mayor." can change as well as the amount of money made. I then need it to fetch how much I made from that Earn and store it as a variable. Sorry I didn't explain this clearer earlier, didn't realise it would be so complex!

 

 

[The Little Guy]

OK... Here are two simple long functions

 

split_string();

and

return_between();

(scroll to bottom of box to see functions working)

<?php
function split_string($string, $delineator, $desired, $type)
    {
    # Case insensitive parse, convert string and delineator to lower case
    $lc_str = strtolower($string);
$marker = strtolower($delineator);
    
    # Return text BEFORE the delineator
    if($desired == BEFORE)
        {
        if($type == EXCL)  // Return text ESCL of the delineator
            $split_here = strpos($lc_str, $marker);
        else               // Return text INCL of the delineator
            $split_here = strpos($lc_str, $marker)+strlen($marker);
        
        $parsed_string = substr($string, 0, $split_here);
        }
    # Return text AFTER the delineator
    else
        {
        if($type==EXCL)    // Return text ESCL of the delineator
            $split_here = strpos($lc_str, $marker) + strlen($marker);
        else               // Return text INCL of the delineator
            $split_here = strpos($lc_str, $marker) ;
        
        $parsed_string =  substr($string, $split_here, strlen($string));
        }
    return $parsed_string;
    }
    
    
function return_between($string, $start, $stop, $type)
    {
    $temp = split_string($string, $start, AFTER, $type);
    return split_string($temp, $stop, BEFORE, $type);
    }
    
    
$text = "You worked at the local bar and made $1456 for the night";
    $worked = return_between($text,"local","and",EXCL);
    $wage = return_between($text,"made","for",EXCL);
    echo $worked.'<br>'.$wage;
?>

 

use the return_between function to return anything between two values (text or html).

- the first value is the full text string

- the second value is the first thing to look for

- the third value is the last next thing to look for

- the fourth value is the weather to include or exclude the value from the returned result

 

if this doesn't make sense ask what you have a question about.

[Wildbug]

I need the script to figure out which Earn I have done. Bearing in mind, the last part: "No taxes were taken as Paris has no mayor." can change as well as the amount of money made. I then need it to fetch how much I made from that Earn and store it as a variable. Sorry I didn't explain this clearer earlier, didn't realise it would be so complex!

 

Well, the original examples were very similar, so a regular expression was easy to write.  If you are trying to match natural language, good freaking luck.  

 

Dollar amounts won't be hard to find as they are always preceeded by a "$", but pulling the noun and verb out of the beginning action might be harder.  If they're always past tense, you could use /You (\w+ed).+?(\$[0-9.,])/i so you would return "walked" & "$91" in the first sentence and "scammed" and "$2714" in the second.

 

What are you doing, writing a script to play an RPG?

[per1os]

Not a fan of that, as it isn't just based around this one text string.. There is a few, basically I'm going to be able to Post these "Earns".

 

The real example would be:

[Note: Only one would be posted at a time]

Case 1:

You walked the beat for ages today - protecting the citizens of Paris and earned yourself a $91 bonus! No taxes were taken as Paris has no mayor.

 

Case 2:

You scammed some poor suckers into your scam - they invested their last $2714 and bought some of your phoney stocks!

 

I need the script to figure out which Earn I have done. Bearing in mind, the last part: "No taxes were taken as Paris has no mayor." can change as well as the amount of money made. I then need it to fetch how much I made from that Earn and store it as a variable. Sorry I didn't explain this clearer earlier, didn't realise it would be so complex!

 

 

 

That can be very complex. The money part is the easy part to determine.

 

<?php
function get_money($string) {
     list($first,$second) = split('$', $test);
     list($money) = split(' ', $test);
     return '$' . $money;
}

$string = "I made $1,065 in 1 days worth of work!";
echo get_money($string);
?>

 

The other part is more work as I can easily say, "You bought $50 worth of stock and lost 10% after earning $100" I am not sure if you get that complex, but the human language is hard to interpret. Because you can have earn in a statement and not actually earned money, "You earned $100 after an initial 10% loss" etc.

 

My suggestion is if you have a list of what you know will be there is store this into an array and have it assigned what it means. Anyways hope this helps you out.