string replace conundrum

[corporateboy]

Okay I have a big amount of text stored in a mysql database, now this text has some dollar amounts in it, what i need to do is convert all dollar amounts in the text to "X" in the output.

Something to do with preg_replace or ereg_replace I think, but I have never used them before so I need help!

---- example of what I want to do

$original_text = "... and we will charge $100.00 for the camp, broken down into our costs of $5 for the bus, $82.50 for each camper and then a profit of $12.50 per ticket sold ...";

the output needs to be
... and we will charge $XXX.XX for the camp, broken down into our costs of $X for the bus, $XX.XX for each camper and then a profit of $XX.XX per ticket sold ...

if possible I would like to make all the dollar amounts link to a page (explaining why they are masked from the public).

Is it possible to search for say the next 10 characters after a $ in a variable and if a number replace with an 'X'? Is there an easier way to do it?

How intensive in processor terms is it to replace it on the fly, should I be doing it before it's saved to the database (and save a copy with the figures in a different table or something).

Any help would be greatly appreciated.

[cgraz]

Why make it so hard? Just use preg_replace

echo preg_replace("([0-9])", "X", $original_text);

Not intensive on the processor at all 😉

Cgraz

[corporateboy]

Won't that replace all numbers including those which aren't prices ?

I need to be able to keep those in.

ie

"Resolution 4.14 we will spend $1,000 on this project"

must become

"Resolution 4.14 we will spend $X,XXX on this project"

[xblue]

that's going to be more difficult. matching the prices should not so such a big problem using a regular expression pattern like
'/\$\d+(.\d+)?/'

but there is no way to find out how many characters the quantifiers matched as far as I know.

some easy approaches:
- use a fixed format as replacement, e.g. always $X.XX
- if your data is somewhat predictable, you could set up arrays for the different formats, like

$patterns = array('/\$\d/', '/\$\d\d/', '/\$\d\d\d/', '/\$\d.\d\d/');
$replacements = array('$X', '$XX', '$XXX', '$X.XX');
$str = preg_replace($patterns, $replacements, $str);

if you'd need it to be more flexible, you might use preg_replace_callback or break it down into 2 steps.

[corporateboy]

thanks I will give that a whirl - it is likely to be only 10 different $ strings ie one of the following (1 = any number)

$1
$1.00
$11
$11.00
$111
$111.00
$1111
$1111.00
$11111
$11111.00

so I run a preg_replace for each of the above strings to replace the numbers with X's right ?

so many slashes in the php replace functions - makes my head hurt!

[xblue]

hi, yes, you could do it either one after the other, or creating arrays and use just one call to preg_replace() as in my example.

the example was slightly off though, you should put the longest patterns first in both cases.

[xblue]

I put up an example for the use of preg_replace_callback. may be overkill, but at least it is fairly concise, and I reduced the total number of backslashes to 2.

function digits_to_x($match)
{
    return preg_replace('/[0-9]/', 'X', $match[0]);
}

$string = preg_replace_callback('/\$[0-9]+(\.[0-9]+)?/', 'digits_to_x', $string);

I noticed that you used a dot as decimal separator in your first example and a comma the next time. if you wanted to match both, you'd need to modify the pattern slightly.

[corporateboy]

here is the latest

start code -------------------------------------

$body_old = "This will cost us $";
if($x)
{
$body_old .= $x;
}
else
{
$body_old .= '5';
}

$body = preg_replace('/\$\d\d\d\d\d.\d\d/','$XXXXX.XX', $body_old);
$body = preg_replace('/\$\d\d\d\d\d/','$XXXXX', $body_old);
$body = preg_replace('/\$\d\d\d\d.\d\d/','$XXXX.XX', $body_old);
$body = preg_replace('/\$\d\d\d\d/','$XXXX', $body_old);
$body = preg_replace('/\$\d\d\d.\d\d/','$XXX.XX', $body_old);
$body = preg_replace('/\$\d\d\d/','$XXX', $body_old);
$body = preg_replace('/\$\d\d.\d\d/','$XX.XX', $body_old);
$body = preg_replace('/\$\d\d/','$XX', $body_old);
$body = preg_replace('/\$\d.\d\d/','$X.XX', $body_old);
$body = preg_replace('/\$\d/','$X', $body_old);

echo"body_old: $body_old<br>";
echo"body: $body<br>";
-----------------------------------------------end code

when I make $x = 2000 I get the following output ->

---

body_old: This will cost us $2000

body: This will cost us $X000

any ideas how those pesky 0's can be made X's ?

[xblue]

this cannot work properly because preg_replace() is returning the modified string, it does not modify the original.

did you check the example in my last post?

[corporateboy]

xblue - I tried the callback script and it worked like a charm !

thanks heaps!

I'm learning a lot.

Still trying to replace all the Ø signs (people are cutting and pasting Word stuff into the system which leaves some strange characters) with a few spaces or non-breaking line spaces with the following code

$body = preg_replace('Ø',' ',$body);

but I get this error msg

Warning: Delimiter must not be alphanumeric or backslash

any ideas?

[xblue]

this is because you forgot the delimiter. pcre patterns must be enclosed in delimiters. often a slash is used, as in the previous examples, but you can use any special character except the backslash (as the error message says).

[jpmoriarty]

saw this thread quite late and have been messing around with it a bit. Havent quite finished, but thought this might be of use:

<?
$body_old = "This will cost us $";
$body_old .= '2000';
$body_old .= " and we're not happy about it! $";
$body_old .= '15,000';
$body_old .= "! Can you belive it!";

$money_str_start = strpos($body_old,'$');
$money_str_end = strpos($body_old,' ',$money_str_start);
for($i=1;$i<($money_str_end-$money_str_start);$i++)
{
    @$replacement .= "X";
}

echo "Money found at position $money_str_start, and ends at $money_str_end <br><br>";
$body = preg_replace('/\$(\d){1,}.\d\d/','$'.$replacement, $body_old);


echo"body_old: $body_old<br>";
echo"body: $body<br>";
?>

there are still a few errors, but thought someone else might like to pick up the baton and run with it as it were!