can't find the right PCRE pattern for a preg_match_all

[greatprovider]

i'm starting with a string such as "Na**3C**6H**5O**7*2H**20"

im attempting to match all **\d+ ...once i can match all the double
asterix \d i intend to wrap the \d in "<sub>" tags for display
purposes.

i have been trying to write the correct pattern for 2 weeks now without
any success...i can get preg_replace() to work, with several simple
patterns but when i use preg_match_all, i either get unintended results
or incomplete matches.

any assistance in matching all "\*{2}\d+" from the above string would
be appreciated.

[Alan Little]

Carved in mystic runes upon the very living rock, the last words of
greatprovider of comp.lang.php make plain:

i'm starting with a string such as "Na**3C**6H**5O**7*2H**20"

im attempting to match all **\d+ ...once i can match all the double
asterix \d i intend to wrap the \d in "<sub>" tags for display
purposes.

i have been trying to write the correct pattern for 2 weeks now without
any success...i can get preg_replace() to work, with several simple
patterns but when i use preg_match_all, i either get unintended results
or incomplete matches.

The expression you have works fine for me. You'll have to give more
information about what's happening. "It doesn't work" isn't much to go
on.

--
Alan Little
Phorm PHP Form Processor
http://www.phorm.com/

[greatprovider]

unformated: Ca(C**2H**3O**2)**2

formated: Ca(C₂H**3O₂)₂

using: preg_match_all('/\*{2}[\d]+/', $formula, $my_arr)

any combination of patterns i try, this is as close as i can get.

Alan Little wrote:

Carved in mystic runes upon the very living rock, the last words of
greatprovider of comp.lang.php make plain:

i'm starting with a string such as "Na**3C**6H**5O**7*2H**20"

im attempting to match all **\d+ ...once i can match all the double
asterix \d i intend to wrap the \d in "<sub>" tags for display
purposes.

i have been trying to write the correct pattern for 2 weeks now without
any success...i can get preg_replace() to work, with several simple
patterns but when i use preg_match_all, i either get unintended results
or incomplete matches.

The expression you have works fine for me. You'll have to give more
information about what's happening. "It doesn't work" isn't much to go
on.

--
Alan Little
Phorm PHP Form Processor
http://www.phorm.com/

[Alan Little]

Carved in mystic runes upon the very living rock, the last words of
greatprovider of comp.lang.php make plain:

Alan Little wrote:

Carved in mystic runes upon the very living rock, the last words of
greatprovider of comp.lang.php make plain:

> i'm starting with a string such as "Na**3C**6H**5O**7*2H**20"
>
> im attempting to match all **\d+ ...once i can match all the double
> asterix \d i intend to wrap the \d in "<sub>" tags for display
> purposes.
>
> i have been trying to write the correct pattern for 2 weeks now
> without any success...i can get preg_replace() to work, with
> several simple patterns but when i use preg_match_all, i either get
> unintended results or incomplete matches.

The expression you have works fine for me. You'll have to give more
information about what's happening. "It doesn't work" isn't much to
go on.

unformated: Ca(C**2H**3O**2)**2

formated: Ca(C₂H**3O₂)₂

using: preg_match_all('/\*{2}[\d]+/', $formula, $my_arr)

any combination of patterns i try, this is as close as i can get.

What is? What's as close as you can get? You still haven't said what
you're getting. And what are you trying to get? What you have for
"formatted", there is no way you're going to get that with
preg_match_all, just by itself. preg_replace might do it for you, but you
said that you have that working. So what exactly are you trying to do
with preg_match_all?

http://www.catb.org/~esr/faqs/smart-questions.html

I just ran your example above, and I get:

Array
(
[0] => Array
(
[0] => **2
[1] => **3
[2] => **2
[3] => **2
)

)

--
Alan Little
Phorm PHP Form Processor
http://www.phorm.com/

[Chung Leong]

greatprovider wrote:

i'm starting with a string such as "Na**3C**6H**5O**7*2H**20"

im attempting to match all **\d+ ...once i can match all the double
asterix \d i intend to wrap the \d in "<sub>" tags for display
purposes.

i have been trying to write the correct pattern for 2 weeks now without
any success...i can get preg_replace() to work, with several simple
patterns but when i use preg_match_all, i either get unintended results
or incomplete matches.

any assistance in matching all "\*{2}\d+" from the above string would
be appreciated.

Why do you need preg_match_all for what you're trying to do?
preg_replace() should be adequate. If not that,
preg_replace_callback().

I am guessing here, but perhaps you're forgetting to capture the
digits?

echo preg_replace('/\*{2}(\d+)/', '_\1',
"Na**3C**6H**5O**7*2H**20");

Appears to produce a correct representation of Sodium Citrate.

[greatprovider]

the formulas are inputted into a database in the format:
"Na**3C**6H**5O**7*2H**20"
i query the database and pass the list of unformated formulas through
this function in a loop.

function format_replace($formula) {
$search = "/(\*{2}[\d]+)/";
$replace = "_\l";
$formatted = preg_replace( $search, $replace, $formula);
return $formatted;
}
the result i get for this is: Na\lC\lH\lO\l*2H\l (where "\l" is
subscripted).

alternatively i wrote this function using preg_match_all(), you may
notice a few other patterns i have commented out, these too failed my
litmus test so to speak.

function format_formula($formula) {
//$search = array ( '/\*{2}[^\d+]*?[\d+\-\s\d+]/' ) -----
((\*{2})[^\d+]*?([\d+])) -----;
//preg_match_all('/(\*\*((\d+)|(\d+\w)|(\d+\s)))/', $formula,
$my_arr); // get the 2-12 subscripted "/(\*{2}([\d]+[\-][\d]+))/"
preg_match_all('/\*{2}[\d]+/', $formula, $my_arr);
var_dump($my_arr);
foreach($my_arr as $key => $val) {
if ($val == NULL || $val == '') {
return;
} else {
for ($i = 0; $i < sizeof($my_arr); $i++) {
for ($j = 0; $j < sizeof($my_arr[$i]); $j++) {
$fixed[$i][$j] = trim($my_arr[$i][$j], "**");
$fixedstr = "_{" . $fixed[$i][$j] . "}";
$formatted = str_replace($my_arr[$i][$j], $fixedstr , $formula);

return $formatted;
}
}
}
}
}


Chung Leong wrote:

greatprovider wrote:

i'm starting with a string such as "Na**3C**6H**5O**7*2H**20"

im attempting to match all **\d+ ...once i can match all the double
asterix \d i intend to wrap the \d in "<sub>" tags for display
purposes.

i have been trying to write the correct pattern for 2 weeks now without
any success...i can get preg_replace() to work, with several simple
patterns but when i use preg_match_all, i either get unintended results
or incomplete matches.

any assistance in matching all "\*{2}\d+" from the above string would
be appreciated.

Why do you need preg_match_all for what you're trying to do?
preg_replace() should be adequate. If not that,
preg_replace_callback().

I am guessing here, but perhaps you're forgetting to capture the
digits?

echo preg_replace('/\*{2}(\d+)/', '_\1',
"Na**3C**6H**5O**7*2H**20");

Appears to produce a correct representation of Sodium Citrate.

[Rik]

greatprovider wrote:

the formulas are inputted into a database in the format:
"Na**3C**6H**5O**7*2H**20"
i query the database and pass the list of unformated formulas through
this function in a loop.

function format_replace($formula) {
$search = "/(\*{2}[\d]+)/";
$replace = "_\l";
$formatted = preg_replace( $search, $replace, $formula);
return $formatted;
}
the result i get for this is: Na\lC\lH\lO\l*2H\l (where "\l" is
subscripted).

Euhm, I think i found the problem after numerous, all working, examples:

Use the NUMBER 1, it's hard in certain fonts to see the difference, but you
write 'l', not'1'.....
$search = '/\*{2}([\d]+)/';
$replace = '_$1';

Using $N instead of \N is preferred.
Single quotes so PHP won't try to match a '$' to an existing variable.

Grtz,
--
Rik Wasmus

[Chung Leong]

greatprovider wrote:

the formulas are inputted into a database in the format:
"Na**3C**6H**5O**7*2H**20"
i query the database and pass the list of unformated formulas through
this function in a loop.

function format_replace($formula) {
$search = "/(\*{2}[\d]+)/";
$replace = "_\l";
$formatted = preg_replace( $search, $replace, $formula);
return $formatted;
}
the result i get for this is: Na\lC\lH\lO\l*2H\l (where "\l" is
subscripted).

You mistyped the code I provided in a couple places. There is no need
for brackets around \d as it's already a class. You want the
parentheses around \d+, not the whole expression, since you only want
the number (and not the two asterisk). And the replacement is \1 not
\l--meaning what's inside the first pair of parentheses.

Just copy-and-paste.

[greatprovider]

hah...that was the trick...thank you all...

one last question...now would anyone mind explaining me how "$1" works?

Rik wrote:

greatprovider wrote:

the formulas are inputted into a database in the format:
"Na**3C**6H**5O**7*2H**20"
i query the database and pass the list of unformated formulas through
this function in a loop.

function format_replace($formula) {
$search = "/(\*{2}[\d]+)/";
$replace = "_\l";
$formatted = preg_replace( $search, $replace, $formula);
return $formatted;
}
the result i get for this is: Na\lC\lH\lO\l*2H\l (where "\l" is
subscripted).

Euhm, I think i found the problem after numerous, all working, examples:

Use the NUMBER 1, it's hard in certain fonts to see the difference, but you
write 'l', not'1'.....
$search = '/\*{2}([\d]+)/';
$replace = '_$1';

Using $N instead of \N is preferred.
Single quotes so PHP won't try to match a '$' to an existing variable.

Grtz,
--
Rik Wasmus

[Rik]

greatprovider wrote:

hah...that was the trick...thank you all...

one last question...now would anyone mind explaining me how "$1"
works?

In a regular expression, you can "capture" pieces that match a pattern with
(). The number after the $ indicates which piece, captures are numbered from
the first opening '('.

For instance:
**567HJK

'/((\*{2})(\d+))/'

$1 will contain the match: '**567';
$2 will contain the match: '**';
$3 will contain the match: '567';

Normally, pieces that have to match a certain regex, but aren't used any
further, don't need (). In some cases, it's necessary for the pattern. In
that case, you could just use the numbered matches you need, discarding the
others (with multiple captures in a regex, it is absolutely not necessary to
use them all). To keep a complex regex more clear, you could also make a
'non-capturing' group by adding ?: after the opening. For instance:
(?:\s+(\d+)) will capture the digits in $1, instead of $2, because the first
parenthesis is told not to capture anything.

Want to learn more about regexes?
http://www.regular-expressions.info/tutorialcnt.html was a big help for me.

Grtz,
--
Rik Wasmus

[greatprovider]

thank you for the explanation..."it all makes sense now..." until i
find another expression....heh

thank you again,
GP

Rik wrote:

greatprovider wrote:

hah...that was the trick...thank you all...

one last question...now would anyone mind explaining me how "$1"
works?

In a regular expression, you can "capture" pieces that match a pattern with
(). The number after the $ indicates which piece, captures are numbered from
the first opening '('.

For instance:
**567HJK

'/((\*{2})(\d+))/'

$1 will contain the match: '**567';
$2 will contain the match: '**';
$3 will contain the match: '567';

Normally, pieces that have to match a certain regex, but aren't used any
further, don't need (). In some cases, it's necessary for the pattern. In
that case, you could just use the numbered matches you need, discarding the
others (with multiple captures in a regex, it is absolutely not necessary to
use them all). To keep a complex regex more clear, you could also make a
'non-capturing' group by adding ?: after the opening. For instance:
(?:\s+(\d+)) will capture the digits in $1, instead of $2, because the first
parenthesis is told not to capture anything.

Want to learn more about regexes?
http://www.regular-expressions.info/tutorialcnt.html was a big help for me.

Grtz,
--
Rik Wasmus