Shell Programming and Scripting

After pouring over my LTKS and Unix in a nutshell, I'm stuck!

I have a large (BMC Report File) that has breaks on DM (district managers).

After a report header, there is a DM header like:
DM: DBP AARON ROBERTS
At the end of each DM break, there is:
** END OF REPORT **

There are also DM's in the report that I am not interested in and must be skipped.

What I grep on is the 3 character DM code, in this instance it is "DBP'. Once I find this line, I write each line into a new file called:
MONTHLY_PLAN_REPORT_XXX where XXX is DBP
until I reach ** END OF REPORT **, after which I move on to the next needed DM in my FOR loop.

I had thought to read each line until I found what I needed and then start all over again. LTKS states on pg 226 (The authors have gone for years without writing a script with read in it). What other way is there?

Please Help!

Jody

One way,

Code:

awk '$2~/DBP|AAA|BBB/{f=1;rep="MONTHLY_PLAN_REPORT_" $2} f{print >> rep}/\*\* END OF REPORT \*\*/{f=0;close(rep)}' BMC_report

Replace the patterns AAA, BBB,... with your actual codes, divided by the pipe | symbol.

What about this one:

Code:

awk '$1=="DM:" && $2 !~ /EXCLUDE|AAA|BBB/{f=1;rep="MONTHLY_PLAN_REPORT_" $2} f{print >> rep}/\*\* END OF REPORT \*\*/{f=0;close(rep)}' BMC_report

What happens if the "EXCLUDE" list is longer than the "INCLUDE" one ...? In absence of a real input, we most of the time rely on guessing.... Either way the OP has now options to choose from...

Thanks guys. I see I have to do a little more awk study. The inclusion list is much smaller than the exclusion list. About 15 out of 75. As soon as I can get logged in at work, I'll let you know how it went.

Thanks again!

Jody

I have the following shell script so far:

for x in DBG DBH DBT DCP DCR DCV DDM DDN DDO DDX DDZ DEA DEN DEO DFR DFZ DGA DHR DIC
do
make_empty_file_func $IMSDATA/monthly_plan_report_$x # this creates a new file if needed or zeros out an existing file
print $header1 >> $IMSDATA/monthly_plan_report_$x
print $header2 >> $IMSDATA/monthly_plan_report_$x

### Fancy awk stuff here

done

Is there a way to pass in $x to the awk script so it will use it instead of hardcoding the $2~/DBP|AAA|BBB/ ?

Thanks

Jody

Always take the short way

Yes there is, see below, but you're still using some hard coding in your command, because you have the codes inside the script.
I'd put all of your codes in a codes_file, one code per line, and update the file when needed. Also awk can overwrite/create files and insert headers itself:

Code:

$ cat reports

#!/bin/ksh

 while read x 
 do 
   awk -v s="$x" 'BEGIN{rep="MONTHLY_REPORT_" s; print "-Header 1-\n-Header 2-" >rep}$1=="DM:" && $2~s{f=1}f{print> rep}/\*\* END OF REPORT \*\*/{f=0}' BMC_report
 done < codes_file

$ cat codes_file
DBG
DBH
DBT
DCP
DBP
DCV
DDM
DGA
DHR
...

Or run the script with all the codes in the command line, which are passed as positional parameters to the command:

Code:

./reports  DBG DBH DBT DCP DBP DCV DDM DGA DHR DIC ...


$cat reports

#!/bin/ksh
# set -x
 
 for x in $*
  do
   awk -v s="$x" 'BEGIN{rep="MONTHLY_REPORT_" s; print "-Header 1-\n-Header 2-">rep}$1=="DM:"&&$2~var{f=1}f{print> rep}/\*\* END OF REPORT  \*\*/{f=0}' BMC_report
  done

Use full paths for the new file names ( /full/path/to/MONTHLY_PLAN...) inside awk , not $IMSDATA/..., if the command is run from a different directory. Also I see
that you don't have too many symbols, so close(file) might not be needed here.