comparing two cobol source file without sequence numbers

[baruch]

HI
How can i use the diff command, to copare two source file
from colume 7.
thx
Barry

 

[Ogzilal]

Hi,

You can't use diff directly, you have some operations to do before to truncate leading sequence numbers.
That is what this sed script is doing.

sed -e 's/^[0-9]*//' prog1.cbl > prog1.cbl.without_seq
sed -e 's/^[0-9]*//' prog2.cbl > prog2.cbl.without_seq

then

diff -bw prog1.cbl.without_seq prog2.cbl.without_seq

After diff, clean the workin files :
rm prog1.cbl.without_seq
rm prog2.cbl.without_seq

Ali

 

[baruch]

Thx Ali
Thats helpfull,

I want to kake in temp area, two other files that contain the seq. numbers, then to attach the orginal seq, to the compareble list.

I wrote a cobol program in dos/pc that do that job.

Any Idea to do that withot cobol program?

Regrds

Barry

 

[Ogzilal]

Hi Barry,

please post a sample of what you are asking for.

Ali

 

[baruch]

Hi Ali and thx again.

here a sample that are made in Dos,
At End the seq. numberes reprinted in the fc output.

1. First like you, i made two temp. files without the first seq. dight, (can be duplicate seq. numberes).

2. I made also, index file that its key is the relative number that poinet to the seq. number.

3. Use the Dos command:

FC /N FC-R.TMP FC-L.TMP >> FC.TMP

4. Restore the oreginal seq.

5. The output is like that.

COMPARING FILES VCHESHHD.OLD AND VCHESHHD.CBL
VCHESHHD.OLD ******
000026     XOPY VETLIB 201001 THRU 201004.
000027     XOPY NEWLIB 000200 THRU 000204.
VCHESHHD.CBL ******
000026     XOPY VETLIB 201001 THRU 201004.
000027     XOPY VETLIB 203700 THRU 203704.
000027     XOPY NEWLIB 000200 THRU 000204.
******
...

I want to to that task in unix, with sed,perel,awk, etc
with as less as code like cobol.

Regards

Barry

 

[Ogzilal]

Hi Barry,

Try this script, I used semething like this in the past y2k migration to include some waiting modifications that are rising after we take a reference version.



# begin of script
#------------------------------------------------
# control number of arguments
#------------------------------------------------
if [ $# -ne 2 ]
then
print -u2 usage : $0 oldfile newfile
exit 0
fi
OLDFILE=$1
NEWFILE=$2

#------------------------------------------------
# be sure arg1 and arg2 are readable files
#------------------------------------------------
if [ ! -f $OLDFILE ]
then
print -u2 "$OLDFILE not found"
exit 1
fi
if [ ! -f $NEWFILE ]
then
print -u2 "$NEWFILE not found"
exit 1
fi

#--------------------------------------------------------------
# create a file identical to odlfile without sequence numbers
#--------------------------------------------------------------
sed -e 's/^ *[0-9]*//' $OLDFILE >$OLDFILE.wseq

#--------------------------------------------------------------
# create a file identical to newfile without sequence numbers
#--------------------------------------------------------------
sed -e 's/^ *[0-9]*//' $NEWFILE >$NEWFILE.wseq

#--------------------------------------------------------------
# compare the 2 files without sequence numbers and keep differences
# in a file
# the diff is made with the switch -C1 wich prints one line before
# and after the lines that are differentes
#--------------------------------------------------------------
diff -bwC1 $OLDFILE $NEWFILE > $OLDFILE.difwseq
#--------------------------------------------------------------
# founded differences are used as patterns in a file for grep in
# original files, sequence numebers will appear in easy way to read
#--------------------------------------------------------------
grep -f$OLDFILE.difwseq $OLDFILE $NEWFILE | tee $OLDFILE.dif

#--------------------------------------------------------------
# make some cleaning
#--------------------------------------------------------------
rm $NEWFILE.wseq $OLDFILE.difwseq $OLDFILE.wseq

# end of script