What is this regex doing? - regex

/^([a-z]:)?\//i
I don't quite understand what the ? in this regex if I had to explain it from what I understood:
Match begin "Group1 is a to z and :" outside ? (which I don't get what its doing) \/ which makes it match / and option /i "case insensitive".
I realize that this will return 0 or 1 not quiet sure why because of the ?
Is this to match directory path or something ?
If I test it:
$var = 'test' would get 0 while $var ='/test'; would get 1 but $var = 'test/' gets 0
So anything that begins with / will get 1 and anything else 0.
Can someone explain this regex in elementary terms to me?

See YAPE::Regex::Explain:
#!/usr/bin/perl
use strict; use warnings;
use YAPE::Regex::Explain;
print YAPE::Regex::Explain->new(qr/^([a-z]:)?\//i)->explain;
The regular expression:
(?i-msx:^([a-z]:)?/)
matches as follows:
NODE EXPLANATION
----------------------------------------------------------------------
(?i-msx: group, but do not capture (case-insensitive)
(with ^ and $ matching normally) (with . not
matching \n) (matching whitespace and #
normally):
----------------------------------------------------------------------
^ the beginning of the string
----------------------------------------------------------------------
( group and capture to \1 (optional
(matching the most amount possible)):
----------------------------------------------------------------------
[a-z] any character of: 'a' to 'z'
----------------------------------------------------------------------
: ':'
----------------------------------------------------------------------
)? end of \1 (NOTE: because you're using a
quantifier on this capture, only the LAST
repetition of the captured pattern will be
stored in \1)
----------------------------------------------------------------------
/ '/'
----------------------------------------------------------------------
) end of grouping
----------------------------------------------------------------------

It matches a lower- or upper case letter ([a-z] with the i modifier) positioned at the start of the input string (^) followed by a colon (:) all optionally (?), followed by a forward slash \/.
In short:
^ # match the beginning of the input
( # start capture group 1
[a-z] # match any character from the set {'A'..'Z', 'a'..'z'} (with the i-modifier!)
: # match the character ':'
)? # end capture group 1 and match it once or none at all
\/ # match the character '/'

? will match one or none of the preceding pattern.
? Match 1 or 0 times
See also: perldoc perlre
Explanation:
/.../i # case insensitive
^(...) # match at the beginning of the string
[a-z]: # one character between 'a' and 'z' followed by a colon
(...)? # zero or one time of the group, enclosed in ()
So in english: Match anything which begins with a / (slash) or some letter followed by a colon followed by a /. This looks like it matches pathnames across unix and windows, e.g.
it would match:
/home/user
and
C:/Applications
etc.

It looks like it is looking for a "rooted" path. It will successfully match any string that either starts with a forward slash (/test), or a drive letter followed by a colon, followed by a forward slash (c:/test).

Specifically, the question mark makes something optional. It applies to the part in parentheses, which is a letter followed by a colon.
Things that will match:
C:/
a:/
/
(That last item above is why the ? is there)
Things that will not match:
C:
a:
ab:/
a/

Related

Regex to capture optional characters

I want to pull out a base string (Wax) or (noWax) from a longer string, along with potentially any data before and after if the string is Wax. I'm having trouble getting the last item in my list below (noWax) to match.
Can anyone flex their regex muscles? I'm fairly new to regex so advice on optimization is welcome as long as all matches below are found.
What I'm working with in Regex101:
/(?<Wax>Wax(?:Only|-?\d+))/mg
Original string
need to extract in a capturing group
Loc3_341001_WaxOnly_S212
WaxOnly
Loc4_34412-a_Wax4_S231
Wax4
Loc3a_231121-a_Wax-4-S451
Wax-4
Loc3_34112_noWax_S311
noWax
Here is one way to do so, using a conditional:
(?<Wax>(no)?Wax(?(2)|(?:Only|-?\d+)))
See the online demo.
(no)?: Optional capture group.
(? If.
(2): Test if capture group 2 exists ((no)). If it does, do nothing.
|: Or.
(?:Only|-?\d+)
I assume the following match is desired.
the match must include 'Wax'
'Wax' is to be preceded by '_' or by '_no'. If the latter 'no' is included in the match.
'Wax' may be followed by:
'Only' followed by '_', in which case 'Only' is part of the match, or
one or more digits, followed by '_', in which case the digits are part of the match, or
'-' followed by one or more digits, followed by '-', in which case
'-' followed by one or more digits is part of the match.
If these assumptions are correct the string can be matched against the following regular expression:
(?<=_)(?:(?:no)?Wax(?:(?:Only|\d+)?(?=_)|\-\d+(?=-)))
Demo
The regular expression can be broken down as follows.
(?<=_) # positive lookbehind asserts previous character is '_'
(?: # begin non-capture group
(?:no)? # optionally match 'no'
Wax # match literal
(?: # begin non-capture group
(?:Only|\d+)? # optionally match 'Only' or >=1 digits
(?=_) # positive lookahead asserts next character is '_'
| # or
\-\d+ # match '-' followed by >= 1 digits
(?=-) # positive lookahead asserts next character is '-'
) # end non-capture group
) # end non-capture group

Regex to fail if multiple matches found

Take the following regex:
P[0-9]{6}(\s|\.|,)
This is designed to check for a 6 digit number preceded by a "P" within a string - works fine for the most part.
Problem is, we need the to fail if more than one match is found - is that possible?
i.e. make Text 4 in the following screenshot fail but still keep all the others failing / passing as shown:
(this RegEx is being executed in a SQL .net CLR)
If the regex engine used by this tool is indeed the .NET engine, then you can use
^(?:(?!P[0-9]{6}[\s.,]).)*P[0-9]{6}[\s.,](?:(?!P[0-9]{6}[\s.,]).)*$
If it's the native SQL engine, then you can't do it with a single regex match because those engines don't support lookaround assertions.
Explanation:
^ # Start of string
(?: # Start of group which matches...
(?!P[0-9]{6}[\s.,]) # unless it's the start of Pnnnnnn...
. # any character
)* # any number of times
P[0-9]{6}[\s.,] # Now match Pnnnnnn exactly once
(?:(?!P[0-9]{6}[\s.,]).)* # Match anything but Pnnnnnn
$ # until the end of the string
Test it live on regex101.com.
or use this pattern
^(?!(.*P[0-9]{6}[\s.,]){2})(.*P[0-9]{6}[\s.,].*)$
Demo
basically check if the pattern exists and not repeated twice.
^ Start of string
(?! Negative Look-Ahead
( Capturing Group \1
. Any character except line break
* (zero or more)(greedy)
P "P"
[0-9] Character Class [0-9]
{6} (repeated {6} times)
[\s.,] Character Class [\s.,]
) End of Capturing Group \1
{2} (repeated {2} times)
) End of Negative Look-Ahead
( Capturing Group \2
. Any character except line break
* (zero or more)(greedy)
P "P"
[0-9] Character Class [0-9]
{6} (repeated {6} times)
[\s.,] Character Class [\s.,]
. Any character except line break
* (zero or more)(greedy)
) End of Capturing Group \2
$ End of string

How to express this in regular expression?

I need to capture all Upper and Lower case character strings other than "FOO" and "BAR". How to do this?
I tried [^(^FOO$)(^BAR$)] but it doesn't work.
Update:
Actually I'm using this in a context, I am concatenating it with another regex
["(\w)+": _this_regex_ ]
For example ["abc":FOO] shouldn't be matched
All other types say ["abc":BAZ] should match
You want a negative look ahead:
\["(\w+)"\s*:\s*(?!FOO\b|BAR\b)(\w+)]
The (\w+) are capturing group, they store the key/value pairs inside variables (I guess that's what you want to do?)
(?!...) is a negative lookahead: it will cause the regex to fail if what's inside matches.
\b is a word-boundary: here it will make the loohahead match (and so fail the regex) only if FOO is followed by a non alphanum character (so ["foo": FOOLISH] will be accepted by the regex)
\s is a short for all type of whitespaces (spaces, tabs, newlines etc)
Demo: http://regex101.com/r/fM3uZ7
What you tried [^...] was a negative character range: it matches any character (and only one character) that's not inside the character range. And keep in mind that inside character ranges only ], ^ and - are special character (so $ means \$ and so on)
Have a try with:
(?i)^(?!.*foo)(?!.*bar)
In action within perl script:
my $re = qr~(?i)^(?!.*foo)(?!.*bar)~;
while(<DATA>) {
chomp;
say (/$re/ ? "OK : $_" : "KO : $_");
}
__DATA__
["abc":FOO]
["abc":BAZ]
output:
KO : ["abc":FOO]
OK : ["abc":BAZ]
Explanation:
The regular expression:
(?i)^(?!.*foo)(?!.*bar)
matches as follows:
NODE EXPLANATION
----------------------------------------------------------------------
(?i) set flags for this block (case-
insensitive) (with ^ and $ matching
normally) (with . not matching \n)
(matching whitespace and # normally)
----------------------------------------------------------------------
^ the beginning of the string
----------------------------------------------------------------------
(?! look ahead to see if there is not:
----------------------------------------------------------------------
.* any character except \n (0 or more times
(matching the most amount possible))
----------------------------------------------------------------------
foo 'foo'
----------------------------------------------------------------------
) end of look-ahead
----------------------------------------------------------------------
(?! look ahead to see if there is not:
----------------------------------------------------------------------
.* any character except \n (0 or more times
(matching the most amount possible))
----------------------------------------------------------------------
bar 'bar'
----------------------------------------------------------------------
) end of look-ahead
----------------------------------------------------------------------

I can't find proper regexp

I have the following file(like this scheme, but much longer):
LSE ZTX
SWX ZURN
LSE ZYT
NYSE CGI
There are 2 words (like i.e. LSE ZTX) in every line with optional spaces and/or tabs at the beginning, at the end and always in between.
Could someone help me to match these 2 words with regexp? Following the example I wish to have LSE in $1 and ZTX in $2 for the first line, SWX in $1 and ZURN in $2 for the second etc.
I have tried something like:
$line =~ /(\t|\s)*?(.*?)(\t|\s)*?(.*?)/msgi;
$line =~ /[\t*\s*]?(.*?)[\t*\s*]?(.*?)/msgi;
I don't know how can I say, that there could be either spaces or tabs (or both of them mixed, so for ex. \t\s\t)
If you want to just match the two first words, the most basic thing is to just match any sequence of characters that are not whitespace:
my ($word1, $word2) = $line =~ /\S+/g;
This will capture the first two words in $line into the variables, if they exist. Note that parentheses are not required when using the /g modifier. Use an array instead if you want to capture all existing matches.
Always two words, you don't need to match the entire line, so your most simple regex would be:
/(\w+)\s+(\w+)/
I think this is what you want
^\s*([A-Z]+)\s+([A-Z]+)
See it here on Regexr, you find the first code of a row in group 1 and the second in group 2. \s is a whitespace character, it includes e.g. spaces, tabs and newline characters.
In Perl it is something like this:
($code1, $code2) = $line =~ /^\s*([A-Z]+)\s+([A-Z]+)/i;
I think you are reading the text file row by row, so you don't need the modifiers s and m, and g is also not needed.
In case the codes are not only ASCII letters, then replace [A-Z] with \p{L}. \p{L} is a Unicode property that will match every letter in every language.
\s includes also tabulation so your regex looks like:
$line =~ /^\s*([A-Z]+)\s+([A-Z]+)/;
the first word is in the first group ($1) and the second in $2.
You can change [A-Z] to whatever's more convenient with your needs.
Here is the explanation from YAPE::Regex::Explain
The regular expression:
(?-imsx:^\s*([A-Z]+)\s+([A-Z]+))
matches as follows:
NODE EXPLANATION
----------------------------------------------------------------------
(?-imsx: group, but do not capture (case-sensitive)
(with ^ and $ matching normally) (with . not
matching \n) (matching whitespace and #
normally):
----------------------------------------------------------------------
^ the beginning of the string
----------------------------------------------------------------------
\s* whitespace (\n, \r, \t, \f, and " ") (0 or
more times (matching the most amount
possible))
----------------------------------------------------------------------
( group and capture to \1:
----------------------------------------------------------------------
[A-Z]+ any character of: 'A' to 'Z' (1 or more
times (matching the most amount
possible))
----------------------------------------------------------------------
) end of \1
----------------------------------------------------------------------
\s+ whitespace (\n, \r, \t, \f, and " ") (1 or
more times (matching the most amount
possible))
----------------------------------------------------------------------
( group and capture to \2:
----------------------------------------------------------------------
[A-Z]+ any character of: 'A' to 'Z' (1 or more
times (matching the most amount
possible))
----------------------------------------------------------------------
) end of \2
----------------------------------------------------------------------
) end of grouping
----------------------------------------------------------------------
With option "Multiline" this Regex:
^\s*(?<word1>\S+)\s+(?<word2>\S+)\s*$
Will give you N matches each containing 2 groups named:
- word1
- word2
^\s*([A-Z]{3,4})\s+([A-Z]{3,4})$
What this does
^ // Matches the beginning of a string
\s* // Matches a space/tab character zero or more times
([A-Z]{3,4}) // Matches any letter A-Z either 3 or 4 times and captures to $1
\s+ // Then matches at least one tab or space
([A-Z]{3,4}) // Matches any letter A-Z either 3 or 4 times and captures to $2
$ // Matches the end of a string
You can use split here:
use strict;
use warnings;
while (<DATA>) {
my ( $word1, $word2 ) = split;
print "($word1, $word2)\n";
}
__DATA__
LSE ZTX
SWX ZURN
LSE ZYT
NYSE CGI
Output:
(LSE, ZTX)
(SWX, ZURN)
(LSE, ZYT)
(NYSE, CGI)
Assuming the spaces at the start of the line are what you use to identify your codes you want, try this:
Split your string up at newlines, then try this regex:
^\s+(\w+\s+){2}$
This will only match lines that start with some space, followed by a (word - some space - word), then end with some space.
# ^ --> String start
# \s+ --> Any number of spaces
# (\w+\s+){2} --> A (word followed by some space)x2
# $ --> String end.
However, if you want to capture the codes alone, try this:
$line =~ /^\s*(\w+)\s+(\w+)/;
# \s* --> Zero or more whitespace,
# (\w+) --> Followed by a word (group #1),
# \s+ --> Followed by some whitespace,
# (\w+) --> Followed by a word (group #2),
This will match all your codes
/[A-Z]+/

cryptic perl expression

I find the following statement in a perl (actually PDL) program:
/\/([\w]+)$/i;
Can someone decode this for me, an apprentice in perl programming?
Sure, I'll explain it from the inside out:
\w - matches a single character that can be used in a word (alphanumeric, plus '_')
[...] - matches a single character from within the brackets
[\w] - matches a single character that can be used in a word (kinda redundant here)
+ - matches the previous character, repeating as many times as possible, but must appear at least once.
[\w]+ - matches a group of word characters, many times over. This will find a word.
(...) - grouping. remember this set of characters for later.
([\w]+) - match a word, and remember it for later
$ - end-of-line. match something at the end of a line
([\w]+)$ - match the last word on a line, and remember it for later
\/ - a single slash character '/'. it must be escaped by backslash, because slash is special.
\/([\w]+)$ - match the last word on a line, after a slash '/', and remember the word for later. This is probably grabbing the directory/file name from a path.
/.../ - match syntax
/.../i - i means case-insensitive.
All together now:
/\/([\w]+)$/i; - match the last word on a line and remember it for later; the word must come after a slash. Basically, grab the filename from an absolute path. The case insensitive part is irrelevant, \w will already match both cases.
More details about Perl regex here: http://www.troubleshooters.com/codecorn/littperl/perlreg.htm
And as JRFerguson pointed out, YAPE::Regex::Explain is useful for tokenizing regex, and explaining the pieces.
You will find the Yape::Regex::Explain module worth installing.
#!/usr/bin/env perl
use YAPE::Regex::Explain;
#...may need to single quote $ARGV[0] for the shell...
print YAPE::Regex::Explain->new( $ARGV[0] )->explain;
Assuming this script is named 'rexplain' do:
$ ./rexplain '/\/([\w]+)$/i'
...to obtain:
The regular expression:
(?-imsx:/\/([\w]+)$/i)
matches as follows:
NODE EXPLANATION
----------------------------------------------------------------------
(?-imsx: group, but do not capture (case-sensitive)
(with ^ and $ matching normally) (with . not
matching \n) (matching whitespace and #
normally):
----------------------------------------------------------------------
/ '/'
----------------------------------------------------------------------
\/ '/'
----------------------------------------------------------------------
( group and capture to \1:
----------------------------------------------------------------------
[\w]+ any character of: word characters (a-z,
A-Z, 0-9, _) (1 or more times (matching
the most amount possible))
----------------------------------------------------------------------
(with ^ and $ matching normally) (with . not
matching \n) (matching whitespace and #
normally):
----------------------------------------------------------------------
/ '/'
----------------------------------------------------------------------
\/ '/'
----------------------------------------------------------------------
( group and capture to \1:
----------------------------------------------------------------------
[\w]+ any character of: word characters (a-z,
A-Z, 0-9, _) (1 or more times (matching
the most amount possible))
----------------------------------------------------------------------
) end of \1
----------------------------------------------------------------------
$ before an optional \n, and the end of the
string
----------------------------------------------------------------------
/i '/i'
----------------------------------------------------------------------
) end of grouping
----------------------------------------------------------------------
UPDATE:
See also: https://stackoverflow.com/a/12359682/1015385 . As noted there and in the module's documentation:
There is no support for regular expression syntax added after Perl version 5.6, particularly any
constructs added in 5.10.
/\/([\w]+)$/i;
It is a regex, and if it is a complete statement, it is applied to the $_ variable, like so:
$_ =~ /\/([\w]+)$/i;
It looks for a slash \/, followed by an alphanumeric string \w+, followed by end of line $. It also captures () the alphanumeric string, which ends up in the variable $1. The /i on the end makes it case-insensitive, which has no effect in this case.
While it doesn't help "explain" a regex, once you have a test case, Damian's new Regexp::Debugger is a cool utility to watch what actually occurs during the matching. Install it and then do rxrx at the command line to start the debugger, then type in /\/([\w]+)$/ and '/r' (for example), and finally m to start the matching. You can then step through the debugger by hitting enter repeatedly. Really cool!
This is comparing $_ to a slash followed by one or more character (case insensitive) and storing it in $1
$_ value then $1 value
------------------------------
"/abcdes" | "abcdes"
"foo/bar2" | "bar2"
"foobar" | undef # no slash so doesn't match
The Online Regex Analyzer deserves a mention. Here's a link to explain what your regex means, and pasted here for the record.
Sequence: match all of the followings in order
/ (slash)
--+
Repeat | (in GroupNumber:1)
AnyCharIn[ WordCharacter] one or more times |
--+
EndOfLine