Interesting 'awk' problem

[Gafling]

I think I need a real 'regular expression' guru here ...

using this data ( = escape char):
(15Y(s1p12v0s0b0T*757540097970*(10U&l6D(s0p11h0s0b4099T

I apply the following 'awk' rule:
/\033.*\*/ { sub(/\033.*\*/,"&quot }

In my experience, this should locate a line with the target data and then
delete the data beginning with the left-most escape character up thru the
FIRST '*'. The target lines are properly located but when the 'sub'
function is run, the data is deleted thru the second '*'. I am sure the
second (escaped) '*' in the replacement function is getting the regular
expression handler confused but I cannot seem to find out how to correct
the substitution so it can work properly.

I would appreciate anyone's insight ...

 

[Astur]

Hello Gafling,

the problem you are facing here is due to the fact that
regular expresions are "greedy" (if they can "eat" more
will do), i.e. the matched string for a pattern will be the
longest possible.

In this case, there is an alternative rule that will do the
work you need. Try with this:

/\033.*\*/ {sub(/[^*]*\*/,""}

the substitution will look for a string of any characters
except * followed by an *, and deletes it.

If you deal regularly with problems like this, take a look
on Perl. It has very strong support for regular expressions.

regards.

Jose Manuel Blanco
BrainBench MVP for Unix admin.