effective regular expression for pattern matching
 

Hello All,

I have been trying to find a proper regular expression to match (rather unmatch) some strings in the files I have.
I have pasted some strings here from a file named "TestFile":
virtual void Handle(const ABC::reportStatusEngaged& ) = 0;
virtual void Handle(const ABC::reportStatusPathOk& ) = 0;
virtual void Handle(const ABC::reportStatusSteeringOverride& ) = 0;
virtual void Handle(const ABC::reportStatusHeadingConverged& ) = 0;
virtual void Handle(const ABC::reportStatusSpeedOK& ) = 0;
virtual void Handle(const ABC::reportStatusHardwareOK& ) = 0;
virtual void Handle(const ABC::reportStatusImplementOk& ) = 0;
virtual void Handle(const ABC::reportStatusDmuOrientationOk& ) = 0;
virtual void Handle(const ABC::reportStatusValveSteering& ) = 0;
virtual void Handle(const ABC::reportStatusValveOnline& ) = 0;
virtual void Handle(const ABC::reportStatusEngageLimits& ) = 0;
virtual void Handle(const ABC::reportStatusElectricSteer& ) = 0;
virtual void Handle(const ABC::reportStatusValveRequiresPowerCycle& ) = 0;
virtual void Handle(const ABC::reportGuidanceDisengageReasonToken& ) = 0;
virtual void Handle(const ABC::reportGuidanceRemoteEngageToken& ) = 0;
virtual void Handle(const ABC::reportPathErrorsToken& ) = 0;

My aim is to find the strings which have no "Token" in them with Token before the "&". When I run the script on this file after the replacement it should not show any matches.
I tried a number of things out of which one is shown below:
sed -ne 's/\(re.*\)[^Token]&/\1Token\&/' TestFile

The trouble is, it seems to skip some strings like:
reportStatusSteeringOverride&
reportStatusHeadingConverged&
The strings containing "OK".

Please help and advise.

Can you post the desired output from that lot, as your description is unclear to me.

Your current regex says, find lines containing re followed by zero or more characters up until not one of 'T','o','k','e' or 'n' followed by &
Also, as you have used '-n' in your sed, you either need to remove it or put a 'p' at the end of the replacement:
You can probably see why you are getting errors. Looking at your 2 examples:

reportStatusSteeringOverride& ... This will not be displayed as the 'e' prior to & is part of the characters of Token
reportStatusHeadingConverged& ... This will be displayed as 'd' is not part of Token

[^Token] means: contains anything but the letters T,o,k,e or n (and does not mean something which does not contain the world Token).

I would suggest you to search for Token& and skip those lines:
sed ' /Token&/n; s/&/Token&/' file

Hi Chris,

The desired output would be to have all those strings mentioned, to have the world "Token"embedded in them before the "&".
For example, the first one: virtual void Handle(const ABC::reportStatusEngaged& ) should become: virtual void Handle(const ABC::reportStatusEngagedToken& )

Hi Grail,
The problem is "converged" is also not being displayed.

Also, could you guys tell me how to use regexp to negative match an entire word?
I tried [^\<Token\<]. But it does not seem to work.

Regards,
rsmitha.

Maybe something like:
If you are using this to edit a file, change the -n to -i and remove the 'p' at the end.

Hi Grail,

I think this works !!

Thank you !!
I was initiallly doing something like this with sed.
sed -n '/re.*Token&/!p' $file
It was giving me the right matches but I did not know how to feed this to the replacement command.
Your reply has given me that. :)