[RESOLVED] Pattern..eek!

Sangre

Hello

I made a script that reads data from a large copy=pasted chunk of data.
it works like a charm, but it screws up with companynames like XS2ALL

because it takes the '2' as a value

So what i need is a command that replaces names with numbers in them with <bad name>, so my script can continue running.

Well, i looked at php.net... went to ereg_replace()
but looking at the pattern discription makes me dizzy!

what a hard to chew on batch of text 🙂

i came up with:

$text = ereg_replace("[^a-zA-Z0-9A-Za-z]", "<bad name> ", $text);

but that entirely deletes all data, so it not working.

My question, how do i use ereg_replace to find "textnumbertext", but doesnt find "texttexttext"?

any of you good with patterns? 😃

Thnx alot!

laserlight

It might be something like:

$text = preg_replace('/\w+\d+\w+/', '<bad name>', $text);

But of course, whether that works or not depends on the exact nature of your data. If you can, why not change the script to work with names that have numeric characters, instead of just replacing those names?

Sangre

Thnx for the pattern 🙂 they arnt easy to build 🙁

Let me explain the thing i do with the large text.

its pasted like this:

1 30463 CableKing 442 33
2 21354 Super Store 214 01445
3 21889 things4you 255 03627

well i need the '30463' and '442' from every entry
the data is copy=pasted tab-seperated, so it containes a million white-spaces when posted.

so first i delete all the too-many spaces with " +", " ", $text
then i delete all the characters a-Z, because i dont need the names
i keep this:

1 30463 442 33
2 21354 214 01445
3 21889 4 255 03627

Then i explode everything into an array, and i use
array['1'] and array['2']
array['5'] and array['6']
array['9'] and array['10']
to get the data i need

But as you can see, the '4' on row 3 screws it up, because there is an array-key too many now 🙂

that is why i need to delete all names with numbers in them.

It would be easy if i could keep the names yes.
but 'cableking' is 1 word, Super Store is 2 word. so when i use

explode ( $text, " ")

i would have again 1 key too many

It would be so much simpler if i could use the tabs, because the data being copy-pasted is tabseperated. But tab's arnt being transfered when action="post"'ed right? they are converted to many spaces.

Many you know a better solution, but your pattern should allready help me alot 🙂
thnx.

laserlight

Basically, your data is of this format:
<numbers><spaces><numbers><spaces><text><spaces><numbers><spaces><numbers><newline>

where <text> can contain whitespace, correct?

Sangre

Correct, and <text> can contain stupid names like all4noth'n.

Ow and there are no enters, its 1 mega big line 🙂.
there are enters when you paste it, but they are lost like the tabs as soon as its send 🙁

edit: bah the pattern you made didnt work, or i used it wrong. (i bet the last 🙂)
The output is now: 1 2 3 4 5 6 7 8 haha 🙂
lets see if i can figure out what code i made is broken

$text = str_replace (".", "", $_REQUEST['text']); // remove .
$text = str_replace (",", "", $text); // remove ,
$text = preg_replace('/\w+\d+\w+/', 'bad name', $text);


$text = ereg_replace("[^0-9]", " ", $text); //remove all but numbers
$text = ereg_replace(" +", " ", $text); // remove too many spaces

$lijst = explode(" ",$text); // create the array

laserlight

Ah, I think something like this might do the trick:

<?php
$text = "1 30463 CableKing 442 33 2 21354 Super Store 214 01445 3 21889 things4you 255 03627";

$data = array();
$pattern = '/(\d+)\s+(\d+)\s+.+?\s+(\d+)\s+(\d+)/';
if ($count = preg_match_all($pattern, $text, $matches, PREG_SET_ORDER)) {
    for ($i = 0; $i < $count; ++$i) {
        array_shift($matches[$i]);
        $data[] = $matches[$i];
    }
}

print_r($data);
?>

Basically, \d+ matches one or more numerals, \s+ matches one or more whitespace characters, and .+? matches one or more characters, but with a non-greedy match. See PCRE.

Sangre

Wow, making those patterns is really a skill 🙂
the rest of the code is some advanced coding as well, but i can quite understand what it does.

The big thing is, IT WORKS 😃

i had to alter 1 thing, remove '.' and ',' from the text first with str_replace. because some numbers are like 2.400.004

but then it works like a charm

your a real help, thank you very much for your time 😃