Help with regular expression

timski72

Hello,

I have a string containing "it is nine o'clock"

I run it through mysqli_real_escape_string so now I have "it is nine o\'clock"

I want to find whole words so I can make an array of words but I want "o'clock" to be one word instead of being split into "o" and "clock".

I also want to avoid anything but valid characters being allowed so I've tried the following:

preg_match_all('/\b[a-z]+\b/i', $words, $matches);

but matches contains
it,is,nine,o,clock.

Is it possible to modify the regular expression to find o'clock as a whole word?

Thanks,
Tim.

Weedpacket

Include the apostrophe in the character range so that it matches ' as well as a-z.

timski72

Well I had already tried the following:

preg_match_all('/\b[a-z']+\b/i', $words, $matches);

but I get a syntax message in the IDE. I'm using phpDesigner 2008.

I tried escaping it too.

preg_match_all('/\b[a-z/']+\b/i', $words, $matches);

Presumably I'm getting the syntax wrong?

Thanks,
Tim.

Derokorian

You escape with the other slash....

preg_match_all('/\b[a-z\']+b/i', $words, $matches);

timski72

Oops thanks for that.

but matches still contain it,is,nine,o,clock.

I am trying to get it,is,nine,o'clock.

Derokorian

Are you doing it before or after running it through mysqli_real_escape_string? If before I don't know why its not working, if after its because there is a \ present in the string.

This:

<?php

$str = "it is nine o'clock";

if( preg_match_all('/\b[a-z\']+\b/i',$str,$matches) ) {
   var_dump($matches);
} else {
   echo 'No matches found! Please check your pattern and try again.';
}

Produces this output:

array
  0 => 
    array
      0 => string 'it' (length=2)
      1 => string 'is' (length=2)
      2 => string 'nine' (length=4)
      3 => string 'o'clock' (length=7)

timski72

I run it through mysqli_real_escape_string before so it does have the \ as you suggested.

So I have "it is nine o\'clock".

I thought I should run it through to protect against sql injection. Is that overkill as preg_match_all is only going to find valid characters anyway?

Thanks,
Tim.

Derokorian

*_real_escape_string is to prepare strings to database queries. There is no reason to run any strings through this function, when the strings aren't being used for this purpose. Further more, strings that do need to be used in a query should only be escaped for the purpose of the query itself. Other functions your script performs should not use the escaped version.

timski72

Thanks, that makes sense.

timski72

well thanks to your help so far I've managed to extract o'clock as a single element of an array rather than two seperate elements. Similarly it's is now a single element too.

Now though mysql doesn't like the following query:

select wordid, notes from words where word = 'it's'

I get the following error:
#1064 - You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'ï»¿ select wordid, notes from words where word = 'it\'s' LIMIT 0, 30' at line 1

timski72

OK I think I've answered my own question. I escaped the query properly this time and it worked fine. Have to keep an eye on my typing ;-)

timski72

Sorry still struggling on this one.

I've got the script working on my local pack but not on the server. I've cleared the cache and the code is definitely present.

To check I changed the code to echo the query to screen and the output is different on my local pack and the server.

Local pack:
select wordid, notes from words where word = 'it\'s'
select wordid, notes from words where word = 'nine'
select wordid, notes from words where word = 'o\'clock'

Server:
select wordid, notes from words where word = 'it'
select wordid, notes from words where word = 's'
select wordid, notes from words where word = 'nine'
select wordid, notes from words where word = 'o'
select wordid, notes from words where word = 'clock'

As you can see the local pack is escaped whereas the server isn't and "it's" has been split into "it" and "s" and "o'clock" into "o" and "clock"

Is there any reason the regular expression would behave differently on my local pack to on a server?

Code snippets for reference:

$words "it's nine o'clock"
preg_match_all('/\b[a-z\']+\b/i', $words, $matches);
global $videoWords;
$videoWords = $matches[0];

and then

foreach ($videoWords as $resultWord) {


    $resultWord = mysqli_real_escape_string($connection, $resultWord);
    $doesWordExistAlready = "select wordid, notes from words where word = '$resultWord'";
    echo $doesWordExistAlready;

Thanks,
Tim.

johanafm

timski72;10995975 wrote:
Is there any reason the regular expression would behave differently on my local pack to on a server?

timski72;10995975 wrote:
Code snippets for reference:
$words "it's nine o'clock"

Since this contains two syntax errors, it's not much of use as a reference since it can't match executable code on either machine.

timski72;10995975 wrote:
global $videoWords;

Since you are using the global keyword, we'd have to see all code from the global keyword down to the second piece of code you showed us. Otherwise anything could happen in between to the global variable $videoWords in between.

So the question is, how do you actually assign the string to $words, and what does $words actually contain if you inspect it by output directly after assignment. Are you really using the same code on both machines and/or if you retrieve anything from a database, do both servers use the same database?