Reg Ex return values

rulian

Can I have a regular expression match a part of a string, and return that part plus the parts that follow until another match occurs?

Meaning this is my string, and I have a list of nouns

$Nouns[]="runs";
$Nouns[]="walks";
$Nouns[]="talks";
$Nouns[]="barks";

//OR 
$NOUNS="/runs|walks|talks|barks/"

$String = "Sally runs with scissors, walks on water, talks like the wind, and barks like a dog";

What I need the reg ex to do, either using arrays or flat strings for comparrions , and return the matching pattern plus what follows, to produce the following results;

$Matches[0]="runs with scissors, "; 
$Matches[1]="walks on water, "; 
$Matches[2]="talks like the wind, and "; 
$Matches[3]="barks like a dog";

I can split the nouns fine via the flat string comparison using A | B, but I cant tell what's being matched.

bradgrafelman

Try this:

$VERBS="runs|walks|talks|barks";
$String = "Sally runs with scissors, walks on water, talks like the wind, and barks like a dog";

preg_match_all("/(?:$VERBS).*?(?=$VERBS|$)/i", $String, $Matches);

EDIT: By the way, those words you're searching for are verbs, not nouns! :p

rulian

bradgrafelman wrote:
EDIT: By the way, those words you're searching for are verbs, not nouns! :p

LOL 😃

thanks for your help, I'll try it out as soon as I can

rulian

that worked perfectly, thank you for the help
However, I'm having trouble wrapping my head around the expression,
/(?:$VERBS).*?(?=$VERBS|$)/i

can you or someone help me understand the logic, or syntax rather
i dont really know ?:$VERBS or ?=$VERBS do

bradgrafelman

A great site to learn about regular expressions is Regular-Expressions.info. I constantly refer back to there when I do some complex regexp's. Anway, on to the fun:

b[/b] -- The (?:...) stuff is a way to group a part of a regexp without actually creating a subset within the match. For example, if I had simply said b[/b], a $Matches[1] array would have been created with the results of this subset. Since there's no reason to have this extra subset of matches, I used the (?:...) syntax to group the pattern, but not actually match it separately. Without grouping it, it would have run into the ".*?" coming up.

.? -- Common in regexp's, this simply says take any character (".") and match it zero or more times ("") but make it non-greedy ("?"). In case you're not familiar with the regexp engine, it is by default a "greedy" engine, meaning it consumes as much of the string as possible. Instead, we want it to stop when it reaches the very next verb so that it doesn't combine all of the phrases into one match.

b[/b] -- Again, I'm using a special syntax here. This time, the '=' sign afther the '?' indicates I'm doing a look-ahead assertion. Basically, I'm saying that this pattern should consume characters until it sees that the next characters it matches would be either $VERBS or the end of the string ($). A look-ahead is needed because we don't want it to include the next verb in this phrase.

/i -- Nothing complex here, just the case-insensitive flag.

Hope this helps!

EDIT: By the way, I forgot to mention this before and in this post: the "." character does not by default match new lines. So, if your string spans multiple lines, the pattern will not work as expected. To fix this and allow the string to span multiple lines, add an 's' (I believe the name for it is the "dot-all modifier") after the 'i' modifier.