Replacing changing/variable pattern in a string with a new pattern

icydash

Hey guys. So I'm trying to write a script that replaces footnote citations in one format with footnote citations in another format.

For example, the text will read: "Then he went down the street.[1] That is where the police found him [2]. They immediately arrested him at the house. [3]"

And I would like the script to find the [ #] strings and replace them with {#} ... so that the string reads:
"Then he went down the street.{1} That is where the police found him {2}. They immediately arrested him at the house. {3}"

Note that there will be double, and even tripple, digit footnotes, so the script has to be able to recognize [##] or [###] and replace the brackets with {} accordingly.

I'm not really sure how to do this, since the "pattern" is constantly changing depending on the number between the brackets. Ideas?

Thanks!

bradgrafelman

One pattern that might suit your needs:

/\[([0-9]+)\]/

or if you want to limit it to up to three digits:

/\[([0-9]{1,3})\]/

and then you could even get fancier and ignore leading zeros (e.g. [0] or [012] wouldn't be matched):

/\[([1-9][0-9]{0,2})\]/

EDIT: Forgot to clarify... the "patterns" above are regular expression patterns, namely of the [man]PCRE[/man] variety. Thus, you'd use one of them with a function like [man]preg_replace/man.

icydash

For some reason I tried that and it doesn't seem to work:

echo preg_replace('/\(([1-9][0-9]{0,2})\)/', 'hi', $content);

$content is a textarea input in which I'm putting in "asdjkasdkjh.[1] askljasjlkdalsjkdsa. [2] askd asd asd ad . [3]"

It should be replacing ever [#] instance with "hi", but it doesn't seem to do anything at all for some reason.

Also, what do I do for the "replacement" pattern in the preg_replace? Because I want it to keep the # value the same, and just replace the [] around the number with {}.

bradgrafelman

Oops - wasn't thinking very clearly. The '(' means the pattern is matching a literal '(' character... however, that should've been a '[' (as it now reads) since we want to match a literal '[' character. (Likewise for ')' and ']'.)

icydash;10998200 wrote:
Also, what do I do for the "replacement" pattern in the preg_replace? Because I want it to keep the # value the same, and just replace the [] around the number with {}.

That's why the number portion of the pattern is surrounded with unescaped ()'s - this is a subpattern, which you can reference in the 'replace' string using \1 or $1 for the first subpattern, \2 or $2 for the second subpattern, etc. See the following manual page for info about subpatterns: [man]regexp.reference.subpatterns[/man].

icydash

Never mind, figured it out!

$pattern = array("/\[([1-9][0-9]{0,2})\]/");
$replacement = array("{\\1}");

$strOut = preg_replace($pattern, $replacement, $content);
echo $strOut;