On Thu, May 27, 2010 at 12:28 AM, Roland Mainz <[email protected]> wrote: > On Thu, May 27, 2010 at 12:17 AM, Jack Schwartz > <[email protected]> wrote: >> Thanks for helping... > [snip] >> # read file into variable "input" >> input="$( < 'zzz' )" >> >> # pick the data, \3 refers to the content matched by the pattern in >> the 3rd ()-bracket pair >> match="${input/~(Elr-g).*([^[:space:]]*ramdisk[^[:space:]]*)([[:space:]]+)([^[:space:]])([[:space:]]+).*/\3}" >> >> if [[ $match != "$input" ]] ; then >> printf 'match is %q\n' "$match" >> else >> print -u2 'no match found.' >> fi >> -- snip -- >> >> The core in this example is the line... >> -- snip -- >> ${input/~(Elr-g).*([^[:space:]]*ramdisk[^[:space:]]*)([[:space:]]+)([^[:space:]])([[:space:]]+).*/\3} >> -- snip -- >> It grabs the value of variable "input" and returns the content which >> matches the extended regular pattern (that does the ~(Elr-g) specify: >> Extended regular pattern ("E") with left ("l") and right ("r")anchors >> and greedy matching ("-g", the greedy mode matches the shortest >> pattern, the default for regular expressions is to match the longest >> character sequence)) in the 3rd bracket pair. If no match is found the >> full value of variable "input" is returned. >> >> Notes: >> 1.I made the example a bit more verbose, AFAIK the [:space:] can be >> replaced with a plain white space assuming /etc/mnttab doesn't use >> both SPACE and TAB in the same file. >> 2. [^ ] means: This pattern matches a single character which is not a >> space (^ means "not") >> 3. * matches zero or more characters and + matches one or more characters >> ---- > [snip] >> This info is good to know. However, for this particular example, I think it >> is clearer to use nawk just because it will be shorter. Alex helped me with >> this already, and it works: > > Erm... I doubt it is shorter because you run a seperate process which > shuffels around a few MB during loading&&starting. It only _looks_ > shorter (but you are right, for those who don't do regex pattern stuff > often the solution using "awk" looks easier (but IMO at an expense of > being an overkill. IMHO heavywheight filters like sed/awk/tr/etc. > should only be used if there are more data to process than at least > 1MB. It's IMO not efficient to move around flies using elephants...)). > >> /usr/bin/nawk '$1 ~ /ramdisk/ {if ($2 ~ /^\/$/) print $2}' \ >> </etc/mnttab >> >> However, I am running into a problem when I try to do something similar, but >> where I pass the match-string in as a variable: >> >> VARPKG=/var/pkg >> varpkg_fstype=($/usr/bin/nawk \ >> '{if ($2 == $VARPKG) print $3}' </etc/mnttab >> >> Problem is that $VARPKG is not getting interpretted into /var/pkg before >> getting passed to nawk. > > You use '...' to pass the awk program and anything within '...' is > passed as plain literal to awk. If you want that $ gets interpreted by > the shell then you must use "..."-style string literals.
Erm... I forgot something. If you use some kind of parameters in "awk" scripts it is recommended to pass them as _arguments_ and extract the values using ARGV[i]. See nawk(1). Otherwise you may hit some interesting issues with quoting when your parameter value contains characters which may be interpretable as "awk" program fragment. ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) [email protected] \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;) _______________________________________________ caiman-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
