[RESOLVED] ereg not picking up my pattern?

[messels]

hey, hey, hey

so i've been working this same script over for a few hours every few days this week and haven't been able to make a single inch's worth of movement forward.

essentially the issue:
i have a db filled with data that has a few fields that need to be split based on a string pattern. if it's easier to do this in straight up SQL then i'm down but would need hlep.

because of the task at hand, i just need to run this "data clean" script once--all the data i'm putting in the db from different files is of an uniform type; it's just not scalable for applications. i put the part of the code that i'm struggling with in a function so it's easier to find. there's something wrong with the ereg() usage but i'm not sure what. i've used similar implementations of ereg with no problems (some in different parts of this same script here).

here's what i have:

<?php
function pattern_check ($pattern_used, $string_used, $setID)
{
echo "---HITS FUNCTION---<br />";
echo "$pattern_used <br />";
echo "$string_used <br /> ";
echo "$setID <br />";
//checks for the pattern, then splits and updates to DB
if (ereg ($pattern_used, $string_used))
{
$parts = ereg ($pattern_used, $string_used, $individual_parts);
$set_name = $individual_parts [0];
$rarity = $individual_parts[1];
echo "set name: $set_name with rarity: $rarity <br />";
}
else
{"echo either it's no string match or you f'd it up. <br />";}
echo "---ends function---<br />";
}//end of function
?>
<?php
if(isset($_POST['go']))
{
include "database.inc.php";
$get_entries = "SELECT * FROM cards";
$get_entries_query = mysql_query ($get_entries);
//start to loop through entries
    while ($results = mysql_fetch_array ($get_entries_query, MYSQL_ASSOC))
    {
    $power_unsplit = $results ['power'];  //power from original db upload
    $CardID = $results ['cardID'];  //cardID associated w/ this entry
    $type_unsplit =$results['type'];  //type from original db upload
    $rarity_unsplit = $results ['rarity'];  //rarity from original upload

//checks for coupled power/toughness and then splits the pulled data into power, slash, toughness...as an array 1,2,3
////////////
if (ereg ("^([[:alnum:]]+)\/([[:alnum:]]+)", $power_unsplit))
{
$parts = (ereg("^([[:alnum:]]+)\/([[:alnum:]]+)", $power_unsplit, $individual_parts));
$power= $individual_parts [1];
$power = trim($power);
$toughness = $individual_parts [2];
$toughness=trim ($toughness);
echo "new power: $power  and new toughness: $toughness   ---- <br />";
//now updates the approriate fields in the db
$update_db = "UPDATE cards SET power ='$power', toughness='$toughness' WHERE cardID='$CardID'";
echo "$update_db<br />";
$update_query = mysql_query ($update_db);
    if (!$update_query)
    echo mysql_error();
    else
    { echo "power and toughness update successful<br />";  }
}//end of pattern check for coupled power/toughness
////////////////
//check to see if there's a - which indicates a couple data pair in the type field, if so split and update
if (ereg ('-', $type_unsplit))
  {
$type_parts = split ('-',$type_unsplit);
$type = $type_parts [0];
$type = trim ($type);
$sub_type = $type_parts[1];
$sub_type = trim ($sub_type);
//update db
$update_type = "UPDATE cards SET type='$type', sub_type='$sub_type' WHERE cardID ='$CardID'";
$update_type_query = mysql_query($update_type);
if (!$update_type_query)
echo mysql_error();
else
{ echo "type and subtype update successful<br />";  }
   }//end of - check  
if (!ereg ('-', $type_unsplit))
{  echo "it's not a creature.<br />";  }
///////////////////
///check to see if there's a space in the rarity and will split
///also string checks against each set as an individual set. (set as an array)
////////////////
////after array of set names and id's, check for pattern in rarity_unpslit
/// if ther'es a match, update db w/ correct mtg_setID and rarity (split)
///////////////////
$set_list = "SELECT * FROM mtg_set_names";
$set_list_query = mysql_query ($set_list);
$array_counter = 0;
while ($set_results = mysql_fetch_array ($set_list_query, MYSQL_ASSOC))
{
$set_names[$array_counter] =$set_results['mtg_set_name'];
$set_dbID [$array_counter] = $set_results['mtg_setID'];
$array_counter++;
}//end of while for set data
$array_counter = 0;
foreach ($set_names as $set_name)
{
$pattern_used = $set_name;
$setID = $set_dbID[$array_counter];
$string_used = $rarity_unsplit;
$pattern_check = pattern_check ($pattern_used, $string_used, $setID);
$array_counter++;
}//end of foreach used to go through set names

}//end of while
//kill mysql connection and end first if
mysql_close();
}//end of $_POST if
?>
</head>

<body>
<?php
if ($admin_message)
{ echo "Admin Message:  $admin_message <br />"; }
elseif ($admin_error_message)
{ echo "Admin Message:  $admin_error_message <br />"; }
?>
This page cleans up power, toughness and then type, subtype.<br />
<form method="post" action="#" >
Click Go to clean that DB<br />
<input type="submit" name="go" value="go" />
</form>
</body>

[Weedpacket]

First of all, you may want to consider using preg_match(); ereg will be moved out of the PHP core in version 6.

In fact, I suspect you might be used to using preg_match() in the first place; '/' isn't a special character when used in ereg() so it doesn't need to be escaped with '\'. It's needed in preg_match() if you start and end the pattern with '/'. Unless you really are using "\/" in the database field.

    if (preg_match ("/^([a-z0-9]+)\/([a-z0-9]+)/i", $power_unsplit)) 

It's also redundant to check the pattern and then make the same check again if it succeeds the first time:

    if (preg_match ("/^([a-z0-9]+)\/([a-z0-9]+)/i", $power_unsplit, $individual_parts)) 

(You don't actually use $parts).

[man]split/man, as noted on that page, would be better handled by [man]explode[/man].
[man]strpos[/man] would be a better function to use to check for "-" than ereg('-').

I think MySQL provides regular expression functions itself, so more (possibly all) of the work can be done there.

[messels]

was a bit sick today so i haven't had a chance to try this out. thanks for the reply though. it's damn appreciated. i'll update on if it works or not.

also, any idea why it's NOT picking up the pattern? it SHOULD, right?

thx!

[Weedpacket]

messels wrote:

it SHOULD, right?

"\/" != "/".

[messels]

Weedpacket;10903337 wrote:

"\/" != "/".

hey, i see. you're actually working on the wrong part of the code. that particular piece works perfectly.

i convereted the specific section of hte code to try out preg_match and preg_split w/ ZERO results. 🙁 here's that one section for you to look over.

function pattern_check ($pattern_used, $string_used, $setID)
{
echo "---HITS FUNCTION---<br />";
echo "looking for:  $pattern_used <br />";
echo "in string:  $string_used <br /> ";
echo "$setID <br />";
//checks for the pattern, then splits and updates to DB (when the matching finally works)
if (preg_split("/^($pattern_used+)/i", $string_used, $matches))
{
print_r ($matches);  //prints array w/ one entry; no pieces!
$set_name = $matches[1];
$rarity = $matches[2];
echo "set name: $set_name with rarity: $rarity <br />";  //echos only blanks
}
echo "---END OF FUNCTION---<br />";
}//end of function

:queasy:

[Weedpacket]

messels wrote:

hey, i see. you're actually working on the wrong part of the code. that particular piece works perfectly.

Well, all you said was that "there's something wrong with the ereg() usage but i'm not sure what". Not a lot to go on: nothing about what's going in, what should be coming out, what is coming out, or what the whole point of the exercise is.

For that pattern_check() function. You don't mean [man]preg_split[/man]. You mean [man]preg_match[/man]. But unless $pattern_used has parentheses in it, $matches can't have a no. 2 element because there is only one pair of parentheses in the pattern. But you know more about that than me.

[nrg_alpha]

I concur with WP. Sounds like you have your wires crossed so-to-speak with regards to preg_split and preg_match (two very different functions).

With regards to delimiters such as /.. if your pattern uses such characters as in say a file path, I tend to use delimiters other than / (my personal fav is #, but you can use any non-alphanumeric, non-whitespace ASCII characters (but not backslash).

So to help avoid the need to escape characters within your pattern that might also be the delimiters, use an odd character for your delimiter that isn't likely to be in the pattern in the first place (perhaps #, or ! or ~ by example). Otherwise, you'll likely have to escape some characters here and there (and should you forget to escape a certain character that needs escaping, you'll have errors.

[messels]

nrg_alpha;10903418 wrote:

I concur with WP. Sounds like you have your wires crossed so-to-speak with regards to preg_split and preg_match (two very different functions).

With regards to delimiters such as /.. if your pattern uses such characters as in say a file path, I tend to use delimiters other than / (my personal fav is #, but you can use any non-alphanumeric, non-whitespace ASCII characters (but not backslash).

So to help avoid the need to escape characters within your pattern that might also be the delimiters, use an odd character for your delimiter that isn't likely to be in the pattern in the first place (perhaps #, or ! or ~ by example). Otherwise, you'll likely have to escape some characters here and there (and should you forget to escape a certain character that needs escaping, you'll have errors.

i actually had preg_match but it wasn't doing it either so i tried preg_split, got sick of it and posted here.

it SOUNDS like i need to actually do a replace on the pattern, insert a deliminter and THEN split the pattern?

to be a bit more specific, the pattern would say "Hello" and the string would read "Hello world" and my goal is to split the hello away from the world so i can use the two strings for different areas of the db i'm setting up.

so, to circle back, do i used a preg_replace to insert a delimiter after the pattern_used or...?

thanks again y'all for the help thus far. it's super appreciated. :o

[nrg_alpha]

messels;10903424 wrote:

...
it SOUNDS like i need to actually do a replace on the pattern, insert a deliminter and THEN split the pattern?

No. I think you misunderstand... you always use delimiters when dealing with preg.. the pattern always starts with an opening delimiter character, say by example /, then your pattern's content, then the closing matching delimiter, /. So the format is:
/........./
But as I mentioned, it doesn't have to be /.. it can be #.......#, or !.........!, or ~.......~, etc...

So you don't split a pattern, you split a string using the pattern.

to be a bit more specific, the pattern would say "Hello" and the string would read "Hello world" and my goal is to split the hello away from the world so i can use the two strings for different areas of the db i'm setting up.

Since you want to split the example string you provided "Hello world" into two words, the character that can be used to split everything in this case is simply a space (between Hello and world). Since you don't need regex to break something up by spaces, you can use explode() instead.

So consider this small sample program:

$str = 'Hello world';
$array = explode(' ', $str);
echo "<pre>".print_r($array, true);

Output (via pre / print_r)

Array
(
    [0] => Hello
    [1] => world
)

You should only use preg_split if the sequence of characters that determine the split are complex enough to require regex.. otherwise explode will do it faster.

so, to circle back, do i used a preg_replace to insert a delimiter after the pattern_used or...?

I explained that all preg patterns are encased in delimiters. I would urge you to have a look into preg via the manual, and have a look at some tutorial sites such as this one by example. If you really want to get on the ball with regex, I would suggest this book. Once you have furthered yourself in your understand with PCRE, it will all become much clearer.

[messels]

well, i'm still not communicating this correctly. i totally get that the pattern needs to have a delimiter of any non-word character. i totally get that. that's not my question at all. (sorry for not being clear enough).

the situation is a bit more complex than the "one space". maybe i'm simplifying it too much?
here's a REAL string:
"shard of alara common}"
i need to separate the "shards of alara" from "common}".

what i've tried:
explode, split, ereg, preg_match, preg_split, preg_match

i simply cannot get the pattern to split from the entire string. i just need the damn thing to be separated into two pieces. 😐

[nrg_alpha]

One possible solution:

$str = 'shard of alara common}';
$arr = preg_split('#(common})#', $str, -1, PREG_SPLIT_DELIM_CAPTURE);
foreach ($arr as $key => &$val){
    $val = trim($val); // trim off any leading / trailing space in entry
    if(empty($val)){
        unset($arr[$key]); // kill entry if emtpy
    }
}
echo "<pre>".print_r($arr, true);

Output:

Array
(
    [0] => shard of alara
    [1] => common}
)

[messels]

appreciate all the advice so far but it's not working. 🙁

things to keep in mind:
$pattern_used rotates through a number of names (shards of alara, conflux, tempest, etc) and is suppose to compare to the string (which is also going to change according to the field pulled from the db).
heres' what part of the outupt looks like so far:
---HITS FUNCTION---
looking for:

Eventide
in string: Eventide Common
92
Array
(
[0] => Eventide Common
)

1---END OF FUNCTION---

here's the code controlling this output (based in part on nrg_alpha's response):

function pattern_check ($pattern_used, $string_used, $setID)
{
echo "---HITS FUNCTION---<br />";
echo "looking for:  $pattern_used <br />";
echo "in string:  $string_used <br /> ";
echo "$setID <br />";
//checks for the pattern, then splits and updates to DB
$array = preg_split ("/($pattern_used)/i", $string_used, -1, PREG_SPLIT_DELIM_CAPTURE);
foreach ($array as $match => &$value)
{
echo "$value<br />";
if (empty($value)) {
	unset($array[$match]); 
	}
}
echo "<pre>".print_r($array);
echo "---END OF FUNCTION---<br />";
}//end of function

[messels]

looking over the output, it looks like "$pattern_used" has a carriage return in there [somehow]. would that affect the pattern search and/or split?

[nrg_alpha]

Well, my solution won't work since you ditched the (common}) as the pattern (I was going by your example string "shard of alara common}"). My solution assumes that there will be common} at the end of the string in question. So I am not sure what's going on, as you on the one hand give a 'real' string example, only to turn around and claim that 'shard of alara' is part of the $pattern_used (which rotates its values)? You're not being consistent in your explanations.

I think you will need to list what your patterns are, and what your string(s) is/are....

[messels]

:o

sorry about all the confusion. i did provide a real string and a real pattern. it was just one of many. anyway, i got sick of messing w/ this and approached the situation differently. here's the code that resolved the situation (i'm def' still open to hearing if there's a better way).

function set_rarity_split ($string_used, $CardID)
{
echo "---HITS FUNCTION---<br />";
	$matches = preg_split ("/([\s])/i", $string_used, -1, PREG_SPLIT_DELIM_CAPTURE);
	foreach ($matches as $match => &$value)
	{
	if (empty($value)) {
	unset($matches[$match]); 
	}
	}
	echo "<pre>".print_r($matches, true);
	$num_words = count ($matches);
	echo "$num_words <br />";
	$array_number = $num_words - 1;
	$rarity = $matches[$array_number];
	echo "$rarity<br />";
	$int=0;
	while ($int < $array_number)
	{
	$edition .= $matches[$int];
	$int++;
	}
	echo "set name: $edition<br />";
	$update_db = "UPDATE cards SET rarity='$rarity', edition ='$edition' WHERE cardID='$CardID'";
	echo "$update_db<br />";
	$update_db_query = mysql_query($update_db);
	if (!$update_db_query)
	echo mysql_error ();

echo "---END OF FUNCTION---<br />";
}//end of function

plus the relevant part from the "non-function" part of hte code:

$get_entries = "SELECT * FROM cards";
$get_entries_query = mysql_query ($get_entries);
//start to loop through entries
    while ($results = mysql_fetch_array ($get_entries_query, MYSQL_ASSOC))
    {
    $power_unsplit = $results ['power'];  //power from original db upload
    $CardID = $results ['cardID'];  //cardID associated w/ this entry
    $type_unsplit =$results['type'];  //type from original db upload
    $rarity_unsplit = $results ['rarity'];  //rarity from original upload
    $number_entries = mysql_num_rows ($get_entries_query);
....
$string_used = $rarity_unsplit;
	$split_set_rarity = set_rarity_split ($string_used, $CardID);

[messels]

caught a small error. need to enclose the preg_split within a ereg() that looks for spaces, otherwise it messes the db up on subsequent script-executions.

here's the code

if(ereg (" ", $string_used))
{
//code from above
...}

[Weedpacket]

Or use [man]strpos[/man]: ereg is being removed from the PHP code in version 6.

But then, this script is supposed to just be run once, yes? 🙂

If there's a better way depends on:

nrg_alpha wrote:

I think you will need to list what your patterns are, and what your string(s) is/are....

But if I read your code right....
Your strings are of the form "edition rarity", where rarity doesn't contain spaces but edition might, and you want to transform that into the two strings "edition" and "rarity".

function set_rarity_split ($string_used, $CardID)
{
echo "---HITS FUNCTION---<br />";

$last_space = strrpos($string_used, ' ');
if($last_space !== false) in case there is no space
{
    $edition = substr($string, 0, $last_space);
    $rarity = substr($string, $last_space+1);
}

echo "set name: $edition<br />";
$update_db = "UPDATE cards SET rarity='$rarity', edition ='$edition' WHERE cardID='$CardID'";
echo "$update_db<br />";
$update_db_query = mysql_query($update_db);
if (!$update_db_query)
echo mysql_error ();

echo "---END OF FUNCTION---<br />";
}//end of function 

[nrg_alpha]

$matches = preg_split ("/([\s])/i", $string_used, -1, PREG_SPLIT_DELIM_CAPTURE);

This will split up everything by spaces in your string (and since you are using \s, this is a short-hand character class for literal spaces, tabs and return carriages) you don't need the i modifier, as this checks for case insensitivity. There is no case sensitivity involved in spaces. You also don't need to surround \s in a character class [...], as I just mentioned, \s is already a character class in and of itself.

I would recommend reading up about PCRE in the manual and tutorial sites such as this one. Having a firm understanding of regex can make things much easier to deal with down the road.