-1

I have a .html with obfuscated class names, something like this class="cd hk aw hs vu me hr". I want to replace only one specific classnames(for example "hk") in the entire html.I need a regex expression that can help me achieve that.

I have found this expression: (?<=class="[^"]*).*(?=")

So I tried the following: (?<=class="[^"]*)(.*hk.*)(?=") but it returns also all the classes within the same body.Result here

It should return the hk word alone

Any help is apreciated.

Wiktor Stribiżew
  • 607,720
  • 39
  • 448
  • 563
Cezar Lacatus
  • 3
  • 2

1 Answers1

1

Mandatory Surgeon General warning: you shouldn't use regex to parse HTML.

In this very limited case, you might use

(<[^>]*class\s*=\s*"[^"]*)\bhk\b([^"]*")

and replace it with \1whatever\2. Then "hk" becomes "whatever" provided it's inside a class inside a HTML tag. The \b specifier will prevent "ohka" from matching and the \s ensures that "class = " and "class=" will be treated the same. You will need a multiline modifier for the case of

<p class
     ="ab hk cd" style="display: block">

You can experiment using https://regex101.com .

Update (clarification):

(<[^>]*class\s*=\s*"[^"]*)\bhk\b([^"]*")

will, indeed, find <body class="ab hk cd" style="x"> and match and replace the sequence <body class="ab hk cd". Note, in the regex, that the first part - the one before "hk" - is parenthesized, and so is the part after.

The "trick" is in the replacement formula, where I specified "\1whatever\2" (in some libraries it could be $1whatever$2, or ${1}whatever${2}, or {$1}whatever{$2} and so on). The meaning is that the content of the first parenthesized group goes into \1, and so on.

So the three groups will be:

'<body class="ab '      captured by '(<[^>]*class\s*=\s*"[^"]*)' into $1
'hk'                    captured by "\bhk\b" and discarded
 cd"                    captured by '([^"]*")' into $2

and the replacement \1whatever\2 will become

 \1                  whatever   \2
 '<body class="ab '  whatever   'cd"'

which is the desired output. The replacement of the first and third groups with themselves feels needless and wasteful, but I know no other means of doing this - regex replacement requires regenerating the whole match, and the "extra" parts are required to correctly frame the match itself.

(You could use this kind of syntax to swap two strings around a pivot, for example).

LSerni
  • 55,617
  • 10
  • 65
  • 107