Using sets to find A-Za-z plus the # and - chars ..?

[PeterJones]

@IanSunlun ,

As is documented, - has special meaning in regex character sets. If you want it to be treated as a literal in a character set, it needs to be either the first or last character in the set.

Compare yours:

to this [A-Za-z#%-]:

or, going back to yours, with the $ in the text file:

vs

… the [#-%] portion of the character set says “characters # through %”, which includes the $ between those, so [#-%] will match # or $ or %. Whereas [#%-] says “match # or % or the literal -”

[PeterJones]

@PeterJones said in Using sets to find A-Za-z plus the # and - chars ..?:

As is documented

Actually, it’s not documented in our character classes section. I will remedy that.

[IanSunlun]

@PeterJones
My search term is not finding the URL in my html page.

html page (its not finding this, but it should):
http://mysitename.net/index.php/New_Video#column-one"

[PeterJones]

@IanSunlun said in Using sets to find A-Za-z plus the # and - chars ..?:

http://mysitename.net/index.php/New_Video#column-one"

Um, no it shouldn’t. New_Video#column-one is more than one character. [A-Za-z%#_-] only matches one character.

I think what you want is http://mysitename.net/index.php/[A-Za-z%#_-]+" , which wants one or more charaters from that set.

Also, I hope you don’t have a URL like http://mysitename.net/index.php/one1#column2

Or http://school.edu/~username/o.n.e.#2 , which is something I might have had back in my university homepage days, lo those two-and-a-half decades ago.

Maybe use http://mysitename.net/index.php/[\w%#.~-]+", since \w encompases the [A-Za-z0-9_] portion, and it adds in the URL-safe characters of . and ~, as well as the # separator and %-encoding-start.

[José Luis Montero Castellanos]

@IanSunlun
Hello :) Try this in Npp: (Just to easily verify that it matches)

Find: [.#\-%]

Inside a character class [set]:

The character # is literal
The character % is literal
The . It is literal (remember that outside equals any character.)
\- The only one that needs an escape sequence using \ .

So:
[A-Za-z#\-%.]
The second hyphen is inside in an escape sequence (preceded by \ ).

Another character that needs escape is ^ because of its negation meaning within the brackets [\^].

[IanSunlun]

@PeterJones Ah, thats seems to work thanks.
Does [\w%#.~-]+ put whatever it matches into ${1} ?

[PeterJones]

This post is deleted!

[PeterJones]

This post is deleted!

[PeterJones]

@IanSunlun said in Using sets to find A-Za-z plus the # and - chars ..?:

Does [\w%#.~-]+ put whatever it matches into ${1} ?

Sorry, when I answered, I had forgotten that you previously said,

(So I need to store pagename in ${1} and bookmark in ${2}.)

Putting the # into either match is not what you want, either. You really need two groups, one before the # and one after.

FIND = http://mysitename.net/index.php/([\w%.~-]+)#([\w%.~-]+)"
will only match if there is a bookmark, and the # will not be inside the ${2} group. If you want the # to be included in ${2}, use http://mysitename.net/index.php/([\w%.~-]+)(#[\w%.~-]+)"

[IanSunlun]

@PeterJones said in Using sets to find A-Za-z plus the # and - chars ..?:

FIND = http://mysitename.net/index.php/([\w%.~-]+)#([\w%.~-]+) "

With the period . inbetween the % and the ~ it did not find:
http://mysitename.net/index.php/New_Video#column-one"
But taking the period out, it did find it.
Whats the thinking behind the period in this context ?

[PeterJones]

@IanSunlun ,

Except for -, order doesn’t matter inside the [] character class. The period is there because New.Video#column-one is also a valid URL ender end-string.

FIND = http://mysitename.net/index.php/([\w%.~-]+)#([\w%.~-]+)"
does match http://mysitename.net/index.php/New_Video#column-one":

[Alan Kilborn]

@PeterJones said in Using sets to find A-Za-z plus the # and - chars ..?:

FIND = http://mysitename.net/index.php/([\w%.~-]+)#([\w%.~-]+) "

Is it worth pointing out that the first two periods here really aren’t periods but rather “match any char”, because they aren’t escaped? Sure, an unescaped . will match a literal period, but it will match other things as well (obviously).

IMO, OP here needs to stop asking forum questions and go off and study regex.

[guy038]

Hello, @peterjones,

In the post below, Peter :

https://community.notepad-plus-plus.org/post/81643

You said :

Actually, it’s not documented in our character classes section. I will remedy that.

Then, regarding the Character Class feature, may be, this part could be added to the Official Notepad++ Documentation : :

If we consider the following CHARACTER CLASS structure :

[.......]
123456789

The POSSIBLE location(s), in order to find the LITERAL character below, are :

LITERAL Character [    :     POSSIBLE at any position, BETWEEN 2 to 8 
                             POSSIBLE at any position, BETWEEN 2 to 8, if PRECEDED with an ANTI-SLASH character
							 
LITERAL Character ]    :     POSSIBLE at position 2 ONLY
                             POSSIBLE at any position, BETWEEN 2 to 8, if PRECEDED with an ANTI-SLASH character
							 
LITERAL Character -    :     POSSIBLE at position 2
                             POSSIBLE at position 8
                             POSSIBLE at any position, BETWEEN 2 to 8, if PRECEDED with an ANTI-SLASH character
							 
LITERAL Character \    :     POSSIBLE at any position, BETWEEN 2 to 8, if PRECEDED with an ANTI-SLASH character

---

Of course, change this layout as you like !

Best Regards,

guy038

[Alan Kilborn]

@guy038

It is rather awkward to express, but I like your idea.

My idea for expression:

• To use a “literal [” in a character class: Use it directly like any other character, e.g. [ab[c]; “escaping” is not necessary (but is permissible), e.g. [ab\\[c]

• To use a “literal ]” in a character class: Directly right after the opening [ of the class notation, e.g. []abc], OR “escaped” at any position, e.g. [\\]abc] or [a\\]bc]

• To use a “literal -” in a character class: Directly as the first or last character in the enclosing class notation, e.g. [-abc] or [abc-], OR “escaped” at any position, e.g. [\-abc] or [a\-bc]

• To use a “literal \” in a character class: Must be doubled (i.e., \\) inside the enclosing class notation, e.g. [ab\\c]

[PeterJones]

@Alan-Kilborn & @guy038 ,

I like those suggestions, especially the way Alan rephrased it: it works much better than my clunky first attempt in the manual, that only included - and was not not very readable.

Thanks.

[Alan Kilborn]

@PeterJones

Maybe my first-of-4 bullet points previously should be moved to be the last-of-4, and changed to:

• To use any other literal character in a character class, just use it directly, i.e., no “escaping” needed

---

Maybe it works well as a 2 column 4 row table, headers:

• Character
• To use it literally in a character class

With those headers, the “cell contents” for column 2 could be appropriately shortened to remove redundant verbiage.

[guy038]

Hi, @peterjones,

BTW, Peter, do you intend to include, in some way, the end part of this post, regarding the Free-space mode, which is in the Notes section ?

https://community.notepad-plus-plus.org/post/81368

---

Also, did you correctly receive, by e-mail, my attached text file, regarding the TextFX features ?

Please, I do not want to stress you, unnecessarily ! Just go at your own pace !

Best Regards

guy038

[Alan Kilborn]

@guy038 said in Using sets to find A-Za-z plus the # and - chars ..?:

do you intend to include, in some way, the end part of this post, regarding the Free-space mode

He already did, see HERE .

[Andrew McP]

@Alan-Kilborn I really admire you guys for figuring out Regular Expressions; I bet you never get lost in real life when you can keep track of the patterns/positions so well, aka good spatial awareness :)

Oh and I like the trick of having - as last character before ]

[Alan Kilborn]

@Andrew-McP said in Using sets to find A-Za-z plus the # and - chars ..?:

I really admire you guys for figuring out Regular Expressions

So if someone says they have “figured out regular expressions”, I pity them. Because it just means they are ripe for an upcoming whipping when a regex misunderstanding of theirs really embarrasses them. :-)

It pays to always be humble when discussing regular expressions with others. :-)

I bet you never get lost

GPS!

I like the trick of having - as last character before ]

Not so much a trick, as a logical place to put it when you realize that anywhere except the first or last position it must form some sort of “range”.