We Keep Coding

Finding multiple groups in Java regex for simple option parser

I need to modify this regex to find multiple group matches:
(?:--)(?<key>[^\s=]+)(?:(?<assign> *[ =] *)(?! --)(?<value>"[^"]*"|\S+))?
In Java:
"(?:--)(?<key>[^\\s=]+)(?:(?<assign> *[ =] *)(?! --)(?<value>\"[^\"]*\"|\\S+))?"
This matches the following correctly:
--key=value
--key=--value
--key value
--flag
--key="--value"
--key "--value"
--key=value --foo=bar
--key=value --foo=bar --flag
But it fails if --flag comes before any other options:
--key=value --flag --foo=bar
I've been trying to modify the negative lookahead between the assign and value capture groups without success so far. The value captured for flag ends up being --foo=bar instead of null.
Any expert recommendations on how to solve this?

I managed to fix the regex. The website https://regexr.com/ was invaluable.
The fixed regex is:
(?<prefix>--)(?<key>[^\s=]+)(?:(?! --)(?<assign> *[ =] *)(?! --)(?<value>"[^"]*"|\S+))?
Here's the Java class and unit test:
https://gist.github.com/kirklund/845baf340a1999a57db9e59e6ba40ce0

Regex exclude lines not containing words but some words

I am using jenkins editable email plugin where i need to use regex to filter logs
I have this regex
regex="(.*)fatal:(.*)"
which basically matches line start with fatal: so that i can grab errors
Example line look like this
fatal: [localhost]: FAILED! => {"attempts": 1, "changed": true, "cmd": "nslookup test1.local",
But i want to grab all fatal erros except which conatin few worda like nslookup.
I want to ignore the above line which has nslookup and all other should be ok

You can do a negative look ahead with a regex. This would work for your example:
^(.*)fatal: ((?!nsookup).)*$
To check for two different lookahead words:
^(.*)fatal: ((?!nslookup)(?!stuff).)*$
The (.*) at the beginning is only required if there can be something (whitespace or other characters) before "fatal:"
You can play with the expression with this link: https://regex101.com/r/ezxA5s/1

Jenkins Console section: What Java regex will trigger on string ERROR but not on string %%ERRORLEVEL%%?

I am using the Jenkins console sections plugin [1] on a windows server. It is excellent in order to make a nice left navbar on my logs.
Positively, I would like any error message to cause a section header, eg;
Assert-PathExstsNotTooLong : ERROR, The path does not exist: E:\P...
...
Oops! Error, please do not do that.
Negatively, I would like to be able to avoid having spelled-out execution templates cause a new section header, eg the below.
[workspace] $ cmd.exe /C " c:\Windows\Microsoft.NET\Framework64\v4.0.30319\MSBuild.exe /p:Configuration=Debug /p:VisualStudioVersion=12.0 "E:\Program Files (x86)\Jenkins\jobs\M.sln"
Using references here on SO [2] and on the tester you recommended [3], I came up with the following, but it is not working?
^(?=(.*([Ee][Rr][Rr][Oo][Rr] ).*))(?!(%%ERRORLEVEL%%))
Using Regex101's amazing tester, with JS flavor, I used the above as input and had these test strings and outputs. The second line of match info perhaps explains my issue but I do not understand it.
test-strings =
help error you should see me
i am %%errorlevel%% again
i am not a section
match-info;
1. `help error you should see me`
2. `error `
Any tips?
thank you!
1.[] ;This plugin uses Java Regex, per its docs ; ; ; ; X.Collapsing Console Sections Plugin - Jenkins - Jenkins Wiki ; ; https://wiki.jenkins-ci.org/display/JENKINS/Collapsing+Console+Sections+Plugin
2.[] ; An example regex on characters, not strings, to avoid; ; ; ; X.java - Regular expression include and exclude special characters - Stack Overflow ; ; Regular expression include and exclude special characters
3.[] ; ; ; ; ; X.Online regex tester and debugger: JavaScript, Python, PHP, and PCRE ; ; https://www.regex101.com/#javascript

(I can't add comments yet, otherwise I'd ask directly, but your example of a spelled-out message template doesn't include the text %%ERRORLEVEL%%, but I assume that it's meant to be a string with %%ERRORLEVEL%% somewhere in the middle of it. Also, as the example isn't quite right, I can't tell exactly what you mean by "not working")
Your problem is that your regex matches ERROR_ (with a space) anywhere in the text, except where the text is exactly %%ERRORLEVEL%%. I think that instead you could write:
^(?=(.*([Ee][Rr][Rr][Oo][Rr])))(?!.*(%%ERRORLEVEL%%)).*
Do you really need to only match ERROR_ (with a space) as opposed to ERROR (whether or not it has a space)? If the former, then you are already excluding %%ERRORLEVEL%%, and you could just use .*(?i:ERROR ).* as the full regex.
The Collapsing Console Sections Plugin uses Java regular expressions, so you can use (?i:ERROR) to match ERROR case-insensitively.
You need a trailing .* before and after your negative-lookahead atom for %%ERRORLEVEL%%, otherwise it will only exclude an exact match
The documentation for the plugin doesn't say whether the pattern has to match a line completely, or if it just matches text within the line. If it matches the line completely, the leading ^ is unnecessary, but won't be doing any harm.
You've got capturing brackets around ERROR and %%ERRORLEVEL%%. If you're not doing anything with that text, then those brackets are unnecessary.
The following regex will match any line with any of ERROR, Error, error etc in it, except lines with any of %%ERRORLEVEL%%, %%ErrorLevel%%, %%errorlevel%% etc.
^(?=.*(?i:ERROR))(?!.*(?i:%%ERRORLEVEL%%)).*

RegEx to match ends with

I need to write regex in java to match domain and subdomain(.domain.com).
Regex should return true for
domain.com
m.domain.com
abc.domain.com
www.domain.com
but returns false for
abcdomain.com
1domain.com
I try to match domain.com and and if preceding character is present then it must be .
I tried various options but it is failing in one or other test cases.

(^|.*?\.)domain\.com
Try this. See demo.
http://regex101.com/r/lB2sH2/1

Try this:
(\.|^)domain.com$
The first part means that there should be a . or nothing
and the $ means, "ends with"

You can try:
(^|\.)domain\.com$
but Java mostly handles only full-line matches, so:
(.+\.)?domain\.com
or you can use the .endWith() method in Java code:
if (domain.equals("domain.com") || domain.endsWith(".domain.com")) {
// do something...
}

I think you want something like this,
(?:\\w+\\.?)?domain\\.com
DEMO

try this regex
\bdomain\.com$
http://rubular.com/r/QG0FtVWtm6

If you don't know what "domain.com" is going to be, this regex below should give you just the subdomain of whatever domain you are looking for. Matches your specifications, including domains that look like abc.net
([a-z]+)(?=\.[a-z]+\.)
DEMO

Extract certain words from predefined sentence using regular expression

I have seemingly simple task, but I have no experience with regular expressions.
I have to parse SMS body with predefined message text, to get out certain information.
Here is one example:
Täname! {FirstName} {LastName} isikukoodiga {PersonCode} on sõlminud EMT Reisikindlustuse lepingu numbriga {PolicyNumber}, mis kehtib alates {CoverStartDate} kell {CoverStartTime} kuni {CoverEndDate} kell {CoverEndTime} (Eesti aja järgi). Hind: {PremiumEur} eurot. Tutvu tingimustega ({Terms}) http://emt.ee/kindlustus. Kahjukäsitluse number +3727330700.
I have to parse out everything that is in curly braces.
I came up with something like this in Java:
public static final String REGEX_CONFIRMATION = "Täname! (.*) (.*) isikukoodiga (.*) on sõlminud EMT Reisikindlustuse lepingu numbriga (.*), mis kehtib alates (.*) kell (.*) kuni (.*) kell (.*) \\(Eesti aja järgi\\). Hind: (.*) eurot. Tutvu tingimustega \\((.*)\\) http://emt.ee/kindlustus. Kahjukäsitluse number \\+3727330700.";
But it parses out only following groups:
{MARIS}, {PLOTS}, {17204046521}, {22414152}, {01.10.2002}, {13:07},
{02.10.2002}, {23:59}.
As you can see {Terms} is missing. And I can't seem to figure out where is the problem?

how about using this pattern?
\{.*?\}

Wouldn't it make more sense to simply use
\{[^{}]*\}
as your regex? In a string, you would need to write that as
"\\{[^{}]*\\}"
Explanation:
\{ # Match an opening brace
[^{}]* # Match any number of characters except braces
\} # Match a closing brace

http://www.java2s.com/Code/Java/Regular-Expressions/Findallmatches.htm
along with the following regex
\{(.*?)\}

Seems correct to me. Use the DOTALL (and in other cases maybe MULTILINE) options. DOTALL can be added as "(?s)Täname!...". Then the ".*" also maps newline chars.
As the prior matches were found this might be it.

Does it work, when You include brackets into your {TERMS} part?
Instead of:
String regex = "...Tutvu tingimustega \\((.*)\\) http://emt.ee/kindlustus. ...";
You could try:
String regex = "...Tutvu tingimustega (.*) http://emt.ee/kindlustus. ...";
OR depending on, what You have in {TERMS} string, You could change _.*_ to _[^)]*_
This way you would find zero to N chars that are not ending bracket.