Shell Programming and Scripting

I have a file which loads procedure/function into ORACLE.Top of the file contains 3 rows which is FILENAME, FILETYPE AND COMMENTS and then the function/procedure etc.

FILENAME:
FILETYPE:
COMMENTS:
<Function/Procedure etc>

The Filename,FILETYPE, COMMENTS will be entered by the dba which will highlight what type of script and why it was used etc.

As part of my loading, I use multiple GREPS to check for FILENAME, FILETYPE, COMMENTS and then load them into logfiles for future viewing. Is there any better way to do this , because If I have 10 lines like this in the future then i may have to write 10 grep commands.

The command I use
echo "Name Of Object - `grep FILENAME $SQLFILE |cut -f2 -d:`" >> $LOGFILE/objects.lst
echo "Object Type - `grep FILETYPE $SQLFILE |cut -f2 -d:`" >> $LOGFILE/objects.lst
... and so on.

Thanks.

grep [-f pattern_file] $SQLFILE
reads one or more patterns from patternfile. Patterns in patternfile are separated by newlines.so, write all the patterns you are searching for in a file and use the above cmd.Is this what you are looking for ?

here's one alternative using awk:

yerra.awk:

Code:

BEGIN{
  tags="FILENAME:|FILETYPE:|COMMENTS:"
  values="Name Of Object|Object Type|Object Comments"

  n=split(tags, arr, "|");
  n=split(values, tmp, "|");
  for(i=1; i <=n; i++) {
     arr[arr[i]]=tmp[i]
     delete arr[i]
  }
}

$1 in arr { 
   match($0,$1); rest=substr($0, RSTART+RLENGTH+1)
   printf("%s - %s\n", arr[$1], rest)
}

nawk -f yerra.awk $SQLFILE >> $LOGFILE/objects.lst

You'll have to update the script "array" if you have other "tags".
Or you can have a config file for the script that can be read in by the script - that will require some modification.

or you can do something like this - not exactly what you have, but...:

Code:

_pat='(FILENAME|FILETYPE|COMMENTS)'
egrep "${_pat}" $SQLFILE | cut -f2 -d : >> $LOGFILE/objects.lst

Not sure if that's what you've been looking for....

The nawk and using the file to check the pattern is good . Can these both be incorporated in the same script rather than 2 scripts one the main script and other script for comparing the patterns ?

not sure what you mean.
the hypothetical call would be:

nawk -f yerra.awk fileWithPatterns2check.txt $SQLFILE >> $LOGFILE/objects.lst

I mean I have to one more script named yerra.awk which I use in my main script to check the pattern which loads the SQLs into the databases.
Can I the code of yerra.awk in this main program, so that I dont have to maintain 2 scripts for it. Since I would implement this in my Production env.

If I need to maintain yerra.awk, a generalized script would be helpful which check the pattern in any file till the newline is reached and generates the o/p.

oh, I see - sure.
It would be something like this in your main shell script:

Code:

.................
.................
nawk '
# here goes the body of the previously posted awk script
' $SQLFILE >> $LOGFILE/objects.lst