Dynamically replace part of a condition block with regex

Question

There is a string in format:

else if($rule=='somerule1')
echo '{"s":1,"n":"name surname"}';
else if($rule=='somerule2')
echo '{"s":1,"n":"another text here"}';
...

"s" can have only number, "n" any text.

In input I have $rule value, and I need to remove the else if block that corresponds to this value. I am trying this:

$str = preg_replace("/else if\(\$rule=='$rule'\)\necho '{\"s\":[0-9],\"n\":\".*\"/", "", $str);

where $str is a string, that contains blocks I mentioned above, $rule is a string with rule I need to remove. But the function returns $str without changes.

What do I do wrong?

For example, script to change "s" value to 1 works nice:

$str = preg_replace("/$rule'\)\necho '{\"s\":[0-9]/", $rule."')\necho '{\"s\":1", $str);

So, probably, I am doing mistake with = symbol, or maybe with space, or with .*.

@mickmackusa what do you mean by mutilating? I have an app and I need to give access to it by some guid. App sends its guid to php script I mentioned, and if the guid(rule) exist, the server will respond 1, if no - 0. I need a script that will add/change/delete rules, that what I am asking about here — Okumaima
– Okumaima, Commented Nov 13, 2020 at 10:26
My thinking is that if you are modifying the contents of a php file like this, there is sure to be a cleaner and more reliable way to adjust rules. The title of this question lacks specificity. According to PSR-12 standards, else if should be elseif and curly braces should be used. — mickmackusa
– mickmackusa ♦, Commented Nov 13, 2020 at 11:50
@mickmackusa yes, this is an php file, but, I have already done my work, with else if, without braces, and all works perfectly — Okumaima
– Okumaima, Commented Nov 14, 2020 at 10:20
I don't question that it works. I am just saying that it is not aligning with the recommended coding standard. @Oku — mickmackusa
– mickmackusa ♦, Commented Nov 14, 2020 at 10:26

mickmackusa · Accepted Answer · 2020-11-14 11:12:52Z

1

The regex pattern can be much less strict, much simpler, and far easier to read/maintain.

You need to literally match the first line (conditional expression) with the only dynamic component being the $rule variable, then match the entire line that immediately follows it.

Code: (Demo)

$contents = <<<'TEXT'
else if($rule=='somerule1')
echo '{"s":1,"n":"name surname"}';
else if($rule=='somerule2')
echo '{"s":1,"n":"another text here"}';
TEXT;

$rule = "somerule1";

echo preg_replace("~\Qelse if(\$rule=='$rule')\E\R.+~", "", $contents);

Output:

else if($rule=='somerule2')
echo '{"s":1,"n":"another text here"}';

So, what have I done? Here's the official pattern demo.

\Q...\E means "treat everything in between these two metacharacters literally"
Then the only character that needs escaping is the first $, this is not to stop it from being interpreted as a end-of-string metacharacter, but as the start of the $rule variable because the pattern is wrapped in double quotes.
The second occurrence of $rule in the pattern DOES need to be interpreted as the variable so it is not escaped.
The \R is a metacharacter which means \n, \r and \r\n
Finally match all of the next line with the "any character" . with a one or more quantifier (+).

answered Nov 14, 2020 at 11:12

mickmackusa♦

49.2k13 gold badges98 silver badges165 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Okumaima Over a year ago

The only problem in Your and @TheFourthBird answer is that it lefts an empty line after replacement and it is not pretty. So I add a \R metacharacter to the end. It always works because (I should mentioned it) in the end, there always is else echo 0; block, so newline after else if will always exist. Thanks.

mickmackusa Over a year ago

Excellent. The extra trailing \R would have been my advice.

The fourth bird · Accepted Answer · 2020-11-13 10:28:01Z

0

The pattern does not match because you have to use a double escape to match the backslash \\\$.

Apart from that, you are not matching the whole line as this part ".*\" stops at the double quote before }';

$str = 'else if($rule==\'somerule1\')
echo \'{"s":1,"n":"name surname"}\';
else if($rule==\'somerule2\')
echo \'{"s":1,"n":"another text here"}\';';

$rule = "somerule1";
$pattern = "/else if\(\\\$rule=='$rule'\)\necho '{\"s\":\d+,\"n\":\"[^\"]*\"}';/";
$str = preg_replace($pattern, "", $str);
echo $str;

Output

else if($rule=='somerule2')
echo '{"s":1,"n":"another text here"}';

Php demo

edited Nov 13, 2020 at 10:28

answered Nov 13, 2020 at 10:12

The fourth bird

165k16 gold badges61 silver badges75 bronze badges

4 Comments

Okumaima Over a year ago

Thanks for feedback, but I didn’t understand about \\n, why then it works in case of $str = preg_replace("/$rule'\)\necho '{\"s\":[0-9]/", $rule."')\necho '{\"s\":1", $str);?

The fourth bird Over a year ago

@Okumaima You are right, it can just be \n. I have updated it.

Okumaima Over a year ago

But $ still should be double quoted?

The fourth bird Over a year ago

@Okumaima Yes because else $ by itself would assert the end of the string.

Collectives™ on Stack Overflow

Dynamically replace part of a condition block with regex

2 Answers 2

2 Comments

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related