Shell Programming and Scripting

I have a set of programs and there coressponding MAPSETs
i tried grep on the all the programs and got the following out put
from this i want to extract only the Program Name and Mapset name {i.e. the word in (' ') after MAPSET }

There or some cases where u have no ( after MAPSET that need to be ignored

input :

A:003150 EXEC CICS RECEIVE MAPSET ('M00DA01') MAP ('DA0001 ') P00DA01
A:004370 EXEC CICS SEND MAPSET ('M00DA01') P00DA01
B:002060 MAPSET ('M0201A')
B:005280 MAPSET ('M0201A')
C:002060 MAPSET ('M0202A')
F: *** M0601A - DEAL INCEPTION MAPSET
G: MAPSET('M0601A')
G: MAPSET('M0601A')
E: MAPSET ('M0602A')
G: MAPSET ('M0601A')
K:011070 MAPSET('M0607A')
L:014510 MAPSET('M0605A')
M:000189 EXEC CICS RECEIVE MAP('M1101A1') MAPSET('M1101A') END-EXEC.
M:000695 EXEC CICS SEND MAP('M1101A1') MAPSET('M1101A')
O: * MAPSET -> M1214A / MAP(S) -> M1214A1 *


Output :


A: M00DA01
B: M0201A
C: M0202A
G: M0601A
E: M0602A
K: M0607A
L: M0605A
M: M1101A

I'm sure there must be a cleaner way to do this, but a sed expert I'm not - I did enjoy dusting off the old regexp reference, though.

If you want to sort on field 1 (the A:, B:, etc), change the 'sort -k2' to 'sort -k1'.

Code:

#! /bin/sh

sed -n "/('.*')/p" | \
 sed -e "s/^\(.\:\).*MAPSET *('\([a-zA-Z0-9]*\)'.*$/\1 \2/g" | \
 sort -k2 -u

Assuming it's named "tmp.sh", run it like './tmp.sh <input-file'

Code:

sed -n -e "s/.*MAPSET *('\([^']*\)'.*/\1/p" | sort -u

i wanted some thing like

A: M00DA01
B: M0201A
C: M0202A
G: M0601A

vino, I knew that could be done in one line. Here's a slightly-updated version of yours that does the same as my earlier attempt:

Code:

 sed -n -e "s/\(.:\).*MAPSET *('\([^']*\)'.*/\1 \2/p" <input-file | sort -k2 -u

one more,

Code:

sed -e "s/.*MAPSET *('//;s/').*//" filename| sed -n '/^M.*/p' | sort -u

you can modify vino's solution :

Code:

sed -n -e "s/\(.*\):.*MAPSET *('\([^']*\)'.*/\1: \2/p" filename | sort -u