How to write regular expression in powershell - regex

I need regular expression in powershell to split string by a string ## and remove string up-to another character (;).
I have the following string.
$temp = "admin#test.com## deliver, expand;user1#test.com## deliver, expand;group1#test.com## deliver, expand;"
Now, I want to split this string and get only email ids into new array object. my expected output should be like this.
admin#test.com
user1#test.com
group1#test.com
To get above output, I need to split string by the character ## and remove sub string up-to semi-colon (;).
Can anyone help me to write regex query to achieve this need in powershell?.

If you want to use regex-based splitting with your approach, you can use ##[^;]*; regex and this code that will also remove all the empty values (with | ? { $_ }):
$res = [regex]::Split($temp, '##[^;]*;') | ? { $_ }
The ##[^;]*; matches:
## - double #
[^;]* - zero or more characters other than ;
; - a literal ;.
See the regex demo

Use [regex]::Matches to get all occurrences of your regular expression. You probably don't need to split your string first if this suits for you:
\b\w+#[^#]*
Debuggex Demo
PowerShell code:
[regex]::Matches($temp, '\b\w+#[^#]*') | ForEach-Object { $_.Groups[0].Value }
Output:
admin#test.com
user1#test.com
group1#test.com

Related

Capture (remove) all double-quotes after colon

I'm trying to clean up a string. An example string:
{
"NodeID": "${NodeID}",
"EventID": "${EventID}"
}
I want to capture all double quotes which occur after the colon, so that the end string will be:
{
"NodeID": ${NodeID},
"EventID": ${EventID}
}
I know that it's JSON, and that technically it is a string in those positions, but they're macros that will be interpreted by a system which generates the actual JSON string and replaces the macros with data, so in my use case this text isn't JSON yet. I can deal with the text line-by-line to make it easier.
I'll be using the regex pattern in both PowerShell and Python.
The closest I've gotten so far have been: (?<=[^*:])("), and (?<=:)(.*)(?<!,)
This is working, but seems incredibly kludgy and inelegant:
$String = '{
"NodeID": "${NodeID}",
"EventID": "${EventID}"
}'
# The Regex to match the text after the colon
[regex]$Regex = '(?<=:)(.*)'
# Splitting each line of the string into an ArrayList element
[System.Collections.ArrayList]$StringArray = $String.Split([string[]][Environment]::NewLine, [StringSplitOptions]::None)
# Declaring an output string
$OutPutString = ''
# Loop through the ArrayList
$i = 1
foreach ($Row in $StringArray) {
# Split each element string at the RegEx match
$RowArray = $Row -split $Regex
[String]$RowString1 = $RowArray[0]
[String]$RowString2 = $RowArray[1]
# Reassemble the element string after replacing the double quotes in the 2nd half
$FullRowString = $RowString1 + $RowString2.Replace('"','')
# If this is the first line in the string, don't add a new line charact in front
if ($i -gt 1) {
$NewLine = "`n"
}
# Reassemble the string
$OutPutString += $NewLine + $FullRowString
$i++
}
$OutPutString
Any better ideas?
👉️ For the regex to be functional as expected, the regex-engine indicated by scripting/programming language is important to know.
Please always add this information as tags besides regex.
Here: powershell, python
Regex to match a JSON text-field and capture the raw-value
Tested on Python, see regex101 demo:
(?<=:\s\s)\"([^\"]*)\"
💡️ Components
To explain the composition of the regex and its working in steps:
(?<=:\s\s): positive look behind ?<=: for 2 white-spaces \s\s
to neglect the field-name also enclosed in double-quotes
\" and \": matching double-quotes before and after the capture group
the unwanted enclosing of the field-value
([^\"]*): capture-group denoted by parentheses surround any non-double-quote character [^\"]*
the wanted raw field-value (string) without enclosing double-quotes
ℹ️ Note:
The character-group [^\"] matches any non (^) double-quote \".
It will start matching at the leading double-quote and stop matching as soon as a double-quote is detected. So the final \" in the regex is optional: It is not required for matching/capturing, but will ensure that each matched field-value is correctly enclosed by double-quotes.
Result
Matching following input lines:
{
"NodeID": "${NodeID}",
"EventID": "${EventID}"
}
Will give the desired raw field-values in group 1 for each match:
e.g.
${NodeID} for the first match
${EventID} for the second match
📚️ Working with JSON in PowerShell
For your context assumed as parsing JSON following related links may be useful:
Microsoft Scripting Blog: Working with JSON data in PowerShell
Related Question: PowerShell parsing JSON
PowerShell Explained: Powershell: The many ways to use regex

perl Regex replace for specific string length

I am using Perl to do some prototyping.
I need an expression to replace e by [ee] if the string is exactly 2 chars and finishes by "e".
le -> l [ee]
me -> m [ee]
elle -> elle : no change
I cannot test the length of the string, I need one expression to do the whole job.
I tried:
`s/(?=^.{0,2}\z).*e\z%/[ee]/g` but this is replacing the whole string
`s/^[c|d|j|l|m|n|s|t]e$/[ee]/g` same result (I listed the possible letters that could precede my "e")
`^(?<=[c|d|j|l|m|n|s|t])e$/[ee]/g` but I have no match, not sure I can use ^ on a positive look behind
EDIT
Guys you're amazing, hours of search on the web and here I get answers minutes after I posted.
I tried all your solutions and they are working perfectly directly in my script, i.e. this one:
my $test2="le";
$test2=~ s/^(\S)e$/\1\[ee\]/g;
print "test2:".$test2."\n";
-> test2:l[ee]
But I am loading these regex from a text file (using Perl for proto, the idea is to reuse it with any language implementing regex):
In the text file I store for example (I used % to split the line between match and replace):
^(\S)e$% \1\[ee\]
and then I parse and apply all regex like that:
my $test="le";
while (my $row = <$fh>) {
chomp $row;
if( $row =~ /%/){
my #reg = split /%/, $row;
#if no replacement, put empty string
if($#reg == 0){
push(#reg,"");
}
print "reg found, reg:".$reg[0].", replace:".$reg[1]."\n";
push #regs, [ #reg ];
}
}
print "orgine:".$test."\n";
for my $i (0 .. $#regs){
my $p=$regs[$i][0];
my $r=$regs[$i][1];
$test=~ s/$p/$r/g;
}
print "final:".$test."\n";
This technique is working well with my other regex, but not yet when I have a $1 or \1 in the replace... here is what I am obtaining:
final:\1\ee\
PS: you answered to initial question, should I open another post ?
Something like s/(?i)^([a-z])e$/$1[ee]/
Why aren't you using a capture group to do the replacement?
`s/^([c|d|j|l|m|n|s|t])e$/\1 [ee]/g`
If those are the characters you need and if it is indeed one word to a line with no whitespace before it or after it, then this will work.
Here's another option depending on what you are looking for. It will match a two character string consisting of one a-z character followed by one 'e' on its own line with possible whitespace before or after. It will replace this will the single a-z character followed by ' [ee]'
`s/^\s*([a-z])e\s*$/\1 [ee]/`
^(\S)e$
Try this.Replace by $1 [ee].See demo.
https://regex101.com/r/hR7tH4/28
I'd do something like this
$word =~ s/^(\w{1})(e)$/$1$2e/;
You can use following regex which match 2 character and then you can replace it with $1\[$2$2\]:
^([a-zA-Z])([a-zA-Z])$
Demo :
$my_string =~ s/^([a-zA-Z])([a-zA-Z])$/$1[$2$2]/;
See demo https://regex101.com/r/iD9oN4/1

regular expression to match multiline text including delimiters

I want to get data between delimiters and include the delimiters in the match.
Example text:
>>> Possible error is caused by the segmentation fault
provided detection report:
<detection-report>
This is somthing that already in the report.
just an example report.
</detection-report>
---------------------------------------------
have a nice day
My current code is:
if($oopsmessage =~/(?<=<detection-report>)((.|\n|\r|\s)+)(?=<\/detection-report>)/) {
$this->{'detection_report'} = $1;
}
It retrieves the following:
This is something that already in the report.
just an example report.
How can i include both the detection-report delimiters?
You can simplify the regex to the following:
my ($report) = $oopsmessage =~ m{(<detection-report>.*?</detection-report>)}s;
Notice I used a different delimiters to avoid the "leaning toothpick syndrome".
The s modifier makes . match newlines.
The parentheses in ($report) force list context, so the match returns all the matching groups. $1 is therefore assigned to $report.
(<detection-report>(?:(?!<\/detection-report>).)*<\/detection-report>)
Try this.Put flags g and s.See demo.
http://regex101.com/r/xT7yD8/18
Just do:
if ($oopsmessage =~ #(<detection-report>[\s\S]+?</detection-report>#) {
$this->{'detection_report'} = $1;
}
or, if you're dreading a file line by line:
while(<$fh>) {
if (/<detection-report>/ .. /<\/detection-report>/) {
$this->{'detection_report'} .= $_;
}
}
Use the below regex to get the data with delimiters.
(<detection-report>[\S\s]+?<\/detection-report>)
Group index 1 contains the string you want.
DEMO
[\S\s] would match one or more space or non-space characters.
/(<detection-report>.*?<\/detection-report>)/gs
You can simplify your regex to the following:
if($oopsmessage =~ m#(<detection-report>.+</detection-report>)#s) {
$this->{'detection_report'} = $1;
}
say $this->{'detection_report'};
Using the modifiers s allows a multiline match where . can be a new line. Using # instead of / means no faffing around with escaping slashes.
Output:
<detection-report>
This is somthing that already in the report.
just an example report.
</detection-report>

Matching numbers for substitution in Perl

I have this little script:
my #list = ('R3_05_foo.txt','T3_12_foo_bar.txt','01.txt');
foreach (#list) {
s/(\d{2}).*\.txt$/$1.txt/;
s/^0+//;
print $_ . "\n";
}
The expected output would be
5.txt
12.txt
1.txt
But instead, I get
R3_05.txt
T3_12.txt
1.txt
The last one is fine, but I cannot fathom why the regex gives me the string start for $1 on this case.
Try this pattern
foreach (#list) {
s/^.*?_?(?|0(\d)|(\d{2})).*\.txt$/$1.txt/;
print $_ . "\n";
}
Explanations:
I use here the branch reset feature (i.e. (?|...()...|...()...)) that allows to put several capturing groups in a single reference ( $1 here ). So, you avoid using a second replacement to trim a zero from the left of the capture.
To remove all from the begining before the number, I use :
.*? # all characters zero or more times
# ( ? -> make the * quantifier lazy to match as less as possible)
_? # an optional underscore
Note that you can ensure that you have only 2 digits adding a lookahead to check if there is not a digit that follows:
s/^.*?_?(?|0(\d)|(\d{2}))(?!\d).*\.txt$/$1.txt/;
(?!\d) means not followed by a digit.
The problem here is that your substitution regex does not cover the whole string, so only part of the string is substituted. But you are using a rather complex solution for a simple problem.
It seems that what you want is to read two digits from the string, and then add .txt to the end of it. So why not just do that?
my #list = ('R3_05_foo.txt','T3_12_foo_bar.txt','01.txt');
for (#list) {
if (/(\d{2})/) {
$_ = "$1.txt";
}
}
To overcome the leading zero effect, you can force a conversion to a number by adding zero to it:
$_ = 0+$1 . ".txt";
I would modify your regular expression. Try using this code:
my #list = ('R3_05_foo.txt','T3_12_foo_bar.txt','01.txt');
foreach (#list) {
s/.*(\d{2}).*\.txt$/$1.txt/;
s/^0+//;
print $_ . "\n";
}
The problem is that the first part in your s/// matches, what you think it does, but that the second part isn't replacing what you think it should. s/// will only replace what was previously matched. Thus to replace something like T3_ you will have to match that too.
s/.*(\d{2}).*\.txt$/$1.txt/;

Regular Expressions: querystring parameters matching

I'm trying to learn something about regular expressions.
Here is what I'm going to match:
/parent/child
/parent/child?
/parent/child?firstparam=abc123
/parent/child?secondparam=def456
/parent/child?firstparam=abc123&secondparam=def456
/parent/child?secondparam=def456&firstparam=abc123
/parent/child?thirdparam=ghi789&secondparam=def456&firstparam=abc123
/parent/child?secondparam=def456&firstparam=abc123&thirdparam=ghi789
/parent/child?thirdparam=ghi789
/parent/child/
/parent/child/?
/parent/child/?firstparam=abc123
/parent/child/?secondparam=def456
/parent/child/?firstparam=abc123&secondparam=def456
/parent/child/?secondparam=def456&firstparam=abc123
/parent/child/?thirdparam=ghi789&secondparam=def456&firstparam=abc123
/parent/child/?secondparam=def456&firstparam=abc123&thirdparam=ghi789
/parent/child/?thirdparam=ghi789
My expression should "grabs" abc123 and def456.
And now just an example about what I'm not going to match ("question mark" is missing):
/parent/child/firstparam=abc123&secondparam=def456
Well, I built the following expression:
^(?:/parent/child){1}(?:^(?:/\?|\?)+(?:firstparam=([^&]*)|secondparam=([^&]*)|[^&]*)?)?
But that doesn't work.
Could you help me to understand what I'm doing wrong?
Thanks in advance.
UPDATE 1
Ok, I made other tests.
I'm trying to fix the previous version with something like this:
/parent/child(?:(?:\?|/\?)+(?:firstparam=([^&]*)|secondparam=([^&]*)|[^&]*)?)?$
Let me explain my idea:
Must start with /parent/child:
/parent/child
Following group is optional
(?: ... )?
The previous optional group must starts with ? or /?
(?:\?|/\?)+
Optional parameters (I grab values if specified parameters are part of querystring)
(?:firstparam=([^&]*)|secondparam=([^&]*)|[^&]*)?
End of line
$
Any advice?
UPDATE 2
My solution must be based just on regular expressions.
Just for example, I previously wrote the following one:
/parent/child(?:[?&/]*(?:firstparam=([^&]*)|secondparam=([^&]*)|[^&]*))*$
And that works pretty nice.
But it matches the following input too:
/parent/child/firstparam=abc123&secondparam=def456
How could I modify the expression in order to not match the previous string?
You didn't specify a language so I'll just usre Perl. So basically instead of matching everything, I just matched exactly what I thought you needed. Correct me if I am wrong please.
while ($subject =~ m/(?<==)\w+?(?=&|\W|$)/g) {
# matched text = $&
}
(?<= # Assert that the regex below can be matched, with the match ending at this position (positive lookbehind)
= # Match the character “=” literally
)
\\w # Match a single character that is a “word character” (letters, digits, and underscores)
+? # Between one and unlimited times, as few times as possible, expanding as needed (lazy)
(?= # Assert that the regex below can be matched, starting at this position (positive lookahead)
# Match either the regular expression below (attempting the next alternative only if this one fails)
& # Match the character “&” literally
| # Or match regular expression number 2 below (attempting the next alternative only if this one fails)
\\W # Match a single character that is a “non-word character”
| # Or match regular expression number 3 below (the entire group fails if this one fails to match)
\$ # Assert position at the end of the string (or before the line break at the end of the string, if any)
)
Output:
This regex will work as long as you know what your parameter names are going to be and you're sure that they won't change.
\/parent\/child\/?\?(?:(?:firstparam|secondparam|thirdparam)\=([\w]+)&?)(?:(?:firstparam|secondparam|thirdparam)\=([\w]+)&?)?(?:(?:firstparam|secondparam|thirdparam)\=([\w]+)&?)?
Whilst regex is not the best solution for this (the above code examples will be far more efficient, as string functions are way faster than regexes) this will work if you need a regex solution with up to 3 parameters. Out of interest, why must the solution use only regex?
In any case, this regex will match the following strings:
/parent/child?firstparam=abc123
/parent/child?secondparam=def456
/parent/child?firstparam=abc123&secondparam=def456
/parent/child?secondparam=def456&firstparam=abc123
/parent/child?thirdparam=ghi789&secondparam=def456&firstparam=abc123
/parent/child?secondparam=def456&firstparam=abc123&thirdparam=ghi789
/parent/child?thirdparam=ghi789
/parent/child/?firstparam=abc123
/parent/child/?secondparam=def456
/parent/child/?firstparam=abc123&secondparam=def456
/parent/child/?secondparam=def456&firstparam=abc123
/parent/child/?thirdparam=ghi789&secondparam=def456&firstparam=abc123
/parent/child/?secondparam=def456&firstparam=abc123&thirdparam=ghi789
/parent/child/?thirdparam=ghi789
It will now only match those containing query string parameters, and put them into capture groups for you.
What language are you using to process your matches?
If you are using preg_match with PHP, you can get the whole match as well as capture groups in an array with
preg_match($regex, $string, $matches);
Then you can access the whole match with $matches[0] and the rest with $matches[1], $matches[2], etc.
If you want to add additional parameters you'll also need to add them in the regex too, and add additional parts to get your data. For example, if you had
/parent/child/?secondparam=def456&firstparam=abc123&fourthparam=jkl01112&thirdparam=ghi789
The regex will become
\/parent\/child\/?\?(?:(?:firstparam|secondparam|thirdparam|fourthparam)\=([\w]+)&?)(?:(?:firstparam|secondparam|thirdparam|fourthparam)\=([\w]+)&?)?(?:(?:firstparam|secondparam|thirdparam|fourthparam)\=([\w]+)&?)?(?:(?:firstparam|secondparam|thirdparam|fourthparam)\=([\w]+)&?)?
This will become a bit more tedious to maintain as you add more parameters, though.
You can optionally include ^ $ at the start and end if the multi-line flag is enabled. If you also need to match the whole lines without query strings, wrap this whole regex in a non-capture group (including ^ $) and add
|(?:^\/parent\/child\/?\??$)
to the end.
You're not escaping the /s in your regex for starters and using {1} for a single repetition of something is unnecessary; you only use those when you want more than one repetition or a range of repetitions.
And part of what you're trying to do is simply not a good use of a regex. I'll show you an easier way to deal with that: you want to use something like split and put the information into a hash that you can check the contents of later. Because you didn't specify a language, I'm just going to use Perl for my example, but every language I know with regexes also has easy access to hashes and something like split, so this should be easy enough to port:
# I picked an example to show how this works.
my $route = '/parent/child/?first=123&second=345&third=678';
my %params; # I'm going to put those URL parameters in this hash.
# Perl has a way to let me avoid escaping the /s, but I wanted an example that
# works in other languages too.
if ($route =~ m/\/parent\/child\/\?(.*)/) { # Use the regex for this part
print "Matched route.\n";
# But NOT for this part.
my $query = $1; # $1 is a Perl thing. It contains what (.*) matched above.
my #items = split '&', $query; # Each item is something like param=123
foreach my $item (#items) {
my ($param, $value) = split '=', $item;
$params{$param} = $value; # Put the parameters in a hash for easy access.
print "$param set to $value \n";
}
}
# Now you can check the parameter values and do whatever you need to with them.
# And you can add new parameters whenever you want, etc.
if ($params{'first'} eq '123') {
# Do whatever
}
My solution:
/(?:\w+/)*(?:(?:\w+)?\?(?:\w+=\w+(?:&\w+=\w+)*)?|\w+|)
Explain:
/(?:\w+/)* match /parent/child/ or /parent/
(?:\w+)?\?(?:\w+=\w+(?:&\w+=\w+)*)? match child?firstparam=abc123 or ?firstparam=abc123 or ?
\w+ match text like child
..|) match nothing(empty)
If you need only query string, pattern would reduce such as:
/(?:\w+/)*(?:\w+)?\?(\w+=\w+(?:&\w+=\w+)*)
If you want to get every parameter from query string, this is a Ruby sample:
re = /\/(?:\w+\/)*(?:\w+)?\?(\w+=\w+(?:&\w+=\w+)*)/
s = '/parent/child?secondparam=def456&firstparam=abc123&thirdparam=ghi789'
if m = s.match(re)
query_str = m[1] # now, you can 100% trust this string
query_str.scan(/(\w+)=(\w+)/) do |param,value| #grab parameter
printf("%s, %s\n", param, value)
end
end
output
secondparam, def456
firstparam, abc123
thirdparam, ghi789
This script will help you.
First, i check, is there any symbol like ?.
Then, i kill first part of line (left from ?).
Next, i split line by &, where each value splitted by =.
my $r = q"/parent/child
/parent/child?
/parent/child?firstparam=abc123
/parent/child?secondparam=def456
/parent/child?firstparam=abc123&secondparam=def456
/parent/child?secondparam=def456&firstparam=abc123
/parent/child?thirdparam=ghi789&secondparam=def456&firstparam=abc123
/parent/child?secondparam=def456&firstparam=abc123&thirdparam=ghi789
/parent/child?thirdparam=ghi789
/parent/child/
/parent/child/?
/parent/child/?firstparam=abc123
/parent/child/?secondparam=def456
/parent/child/?firstparam=abc123&secondparam=def456
/parent/child/?secondparam=def456&firstparam=abc123
/parent/child/?thirdparam=ghi789&secondparam=def456&firstparam=abc123
/parent/child/?secondparam=def456&firstparam=abc123&thirdparam=ghi789
/parent/child/?thirdparam=ghi789";
for my $string(split /\n/, $r){
if (index($string,'?')!=-1){
substr($string, 0, index($string,'?')+1,"");
#say "string = ".$string;
if (index($string,'=')!=-1){
my #params = map{$_ = [split /=/, $_];}split/\&/, $string;
$"="\n";
say "$_->[0] === $_->[1]" for (#params);
say "######next########";
}
else{
#print "there is no params!"
}
}
else{
#say "there is no params!";
}
}