Regex to match decimal point but not .html

Question

I have urls in this format:-

/scan/anything/se=hello-world/se=word.html
/scan/anything/se=hello-world/se=1.5/
/scan/anything/se=temp-2.5/se=1.5.html

I'm trying to match word characters after each se= plus the dash and decimal and capture them.

The regex I have come up with is this:-

  ^/scan/.*?se=([\w-.]*)/?(?:se=)([\w-.]*)/?(?:.html)?

Because I have added a dot(.) in the character class to match the decimal point it also matches .html so captures word.html and 1.5.html rather than just "word" and "1.5" from urls 1 and 3, how can I stop it matching .html I've tried various negations but none seem to work.

Desired output:

• hello-world and word
• hello-world and 1.5
• temp-2.5 and 1.5

Answer 1

You want to use a negative character class like this combined with a positive look ahead, which doesn't count as part of the capturing group:

se=([^/]+)/se=((?:[^/]+)(?=\.html)|[^/]+)

That way you capture all non / up to the next /

Here is a little example in Python:

import re

thelist = [
"/scan/anything/se=hello-world/se=word.html",
"/scan/anything/se=hello-world/se=1.5/",
"/scan/anything/se=temp-2.5/se=1.5.html",
]

regex = "se=([^/]+)/se=((?:[^/]+)(?=\.html)|[^/]+)"

for item in thelist:
    thematch = re.search(regex, item)
    print(thematch.group(1))
    print(thematch.group(2))
    print("------------")

results:

hello-world
word
------------
hello-world
1.5
------------
temp-2.5
1.5
------------

http://regex101.com is a nice little site to play around with this kind of stuff if you need to tweak a regex

Answer 2

I suggest this regex :

se=((?:[\w-.]+)(?=\.html)|[\w-.]+)

See this demo.

This will match any word that can contain - or . until a potential .html (it will stop right before .html if any).

Edit :

The above regex won't capture .html even if it's inside the URL, like at the end of a parameter. For example, this is what would be captured in this case :

/scan/anything/se=hello-world.html/se=word.html
                              ^^^^^^^^^^^              ^^^^^

So if you want to capture everything but the very last .html, you'd have to add an end of string character $ :

se=((?:[\w-.]+)(?=\.html$)|[\w-.]+)

See this second demo.

Edit 2 :

In the light of the information provided by OP's comment down here, this regex would be more appropriate to make URL redirection :

^\/scan\/anything\/se=([\w-.]+)\/se=((?:[\w-.]+)(?=\.html)|[\w-.]+)

See this demo.

This will capture both se parameters in $1 and $2 respectively for each URL while still matching the same inputs as the above regular expressions.