Regex - matching string termination?

[Kadence]

What would one do to match the termination of a string?

For instance

$string = "A sentence. More words. Dot.End.";
$string = preg_replace("#(\.[\sE])#", "[INS]$1", $string);
echo $string; 

Results in:

A sentence[INS]. More words[INS]. Dot[INS].End.

How would one match the end of the string as well as "\s" and "E", so that it becomes:

A sentence[INS]. More words[INS]. Dot[INS].End[INS].

[decowski]

$string = "A sentence. More words. Dot.End.";

$string = preg_replace("#((\.[\sE])|\.$)#", "[INS]$1", $string);

echo $string; 

[Salathe]

Whilst decowski's regex would work just fine, here's a regex which more literally describes what you're after: match a dot, only when followed by either whitespace or the capital E character or the end of the string and replace that dot with "[INS].". It makes uses of a lookahead assertion to catch only the dots that you want.

$string = preg_replace("#\.(?=\s|E|$)#", "[INS].", $string); 

Off Topic:


If you are unfamiliar with the syntax or concept of lookahead assertions (in this case, to match dots followed by specific characters or points in the string) then a helpful introduction can be found in Lookahead and Lookbehind Zero-Width Assertions.

A more basic expression with which you might be more familiar would be #\.(\s|E|$)# with the associated replacement string [INS]$0.

[nrg_alpha]

$string = preg_replace("#\.(?=\s|E|$)#", "[INS].", $string); 

Alternatively (yes, it's a minor alternative):

$string = preg_replace("#(?<=(?:\s|E\$)\.)#", "[INS]", $string); 

At least, it seems to yield the same results. Splitting hairs in the grand scheme of things admittedly.

[Kadence]

Thanks a lot guys

I had originally tried this pattern

#(\.[\sE$])#

But that didn't work so I had thought $ wasn't the right thing to use - but I guess it's perhaps the brackets taking the $ as a literal?

[Salathe]

Alternatively (yes, it's a minor alternative):

$string = preg_replace("#(?<=(?:\s|E\$)\.)#", "[INS]", $string); 

At least, it seems to yield the same results. Splitting hairs in the grand scheme of things admittedly.

I'm not sure how you tested this but it really should not yield the same results. It is looking for a point in the string with either of the following before it: a)a whitespace character followed by a dot, b)capital letter E followed by the end of the string followed by a dot (not possible!).

I guess it's perhaps the brackets taking the $ as a literal?

Absolutely; the $ character has no special meaning within a character class [$…].

[nrg_alpha]

Yeah, I retested and it doesn't work.. I'm not sure what I tested previously.. perhaps a previous result from a previous example that I mistook as the current one or something.. shakes head..