HI Guys

The bash script below runs on a daily basis , i have issue when it comes to the end of the month. in short when i am on the 1 of the next month the script cant find the file which is 30 or 31 file of the previous month.

if anyone can help me tune the script to meet my everyday requirements would be great.


#!/bin/bash
##########################################
############ CONSTANTS###################
#########################################
#06-Feb-2012.txt
#FILE_NAME="`date +%d`-`date +%b`-`date +%Y`".txt
FILE_NAME="`date --date="1 days ago" +%d`-`date +%b`-`date +%Y`".txt
NEWFILE1="`date --date="1 days ago" +%d`-`date +%b`-`date +%Y`"Part1.txt
NEWFILE2="`date --date="1 days ago" +%d`-`date +%b`-`date +%Y`"Part2.txt
SuffixName=$(date +%d%b%Y)
PATH=/home/cruser00/hand/Data
LOG=/var/log/cesar_file.log
DEST=/home/cruser00/handpunch_SHJ
WC=/usr/bin/wc
AWK=/usr/bin/awk
CUT=/usr/bin/cut
BC=/usr/bin/bc
SED=/bin/sed
NOFLINES=$($WC -l $PATH/$FILE_NAME | $CUT -d' ' -f1 | $AWK '{print $1}')
#echo $FILE_NAME
echo "----------start of log for $SuffixName--------------" >> $LOG

if [ -f "$PATH/$FILE_NAME" ];
then
echo "File $FILE_NAME exists" >> $LOG
echo "$FILE_NAME has $NOFLINES number of lines " >> "$LOG"
b=`echo "$NOFLINES % 2"|$BC`
if [ $b -eq 0 ]
then
echo "Given Number is Even "
SPLIT=$(echo "$NOFLINES / 2" | $BC )
SPLITN=$(echo "$SPLIT +1" | $BC)
$SED -n "1,${SPLIT}p" "$PATH/$FILE_NAME" > "$DEST/$NEWFILE1"
$SED -n "${SPLITN},${NOFLINES}p" "$PATH/$FILE_NAME" > "$DEST/$NEWFILE2"
echo " line number 1 -- $SPLIT from $PATH/$FILE_NAME transfered to $DEST/$NEWFILE1 " >> $LOG
echo " line number $SPLITN -- $NOFLINES from $PATH/$FILE_NAME transfered to $DEST/$NEWFILE2 " >> $LOG

echo "----------End of log for $SuffixName--------------" >> $LOG
exit
fi
echo " Given Number is odd"
NEW_NOFLINES=$(echo "$NOFLINES -1" | $BC )
echo $NEW_NOFLINES
SPLIT=$(echo "$NOFLINES / 2" | $BC )
echo " SPLIT = $SPLIT"
SPLITN=$(echo "$SPLIT +1" | $BC)
ADD=$(echo "$NEW_NOFLINES +1" | $BC )
$SED -n "1,${SPLIT}p" "$PATH/$FILE_NAME" > "$DEST/$NEWFILE1"
$SED -n "${SPLITN},${ADD}p" "$PATH/$FILE_NAME" > "$DEST/$NEWFILE2"
echo " line number 1 -- $SPLIT from $PATH/$FILE_NAME transfered to $DEST/$NEWFILE1 " >> $LOG
echo " line number $SPLITN -- $ADD from $PATH/$FILE_NAME transfered to $DEST/$NEWFILE2 " >> $LOG
echo "----------End of log for $SuffixName--------------" >> $LOG
else
echo "File $FILE_NAME does not exists" >> $LOG
echo "----------End of log for $SuffixName--------------" >> $LOG
exit 1
fi

Hi,

the code you posted is extremely difficult to read (IMHO). I think you're more likely to get a useful response if you trim your script to be as simple as possible but still demonstrating the bug/problem. Also you should place it in [code] tags so that fixed width fonts are used and proper indenting is preserved.

Evo2.

1 members found this post helpful.

first, you do not need several date commands:
instead of this:
#FILE_NAME="`date +%d`-`date +%b`-`date +%Y`".txt
you can use:
#FILE_NAME="`date +%d-%b-%Y`".txt
(as an example)

I assume the problem is that you have not used it properly:
FILE_NAME="`date --date="1 days ago" +%d`-`date <you should put 1 day ago here too> +%b`-`date <you should put 1 day ago here too> +%Y`".txt
but would be much better:
FILE_NAME="`date --date="1 days ago" +%d-%b-%Y`".txt
and even better to set:

I think self-documenting code is easier to read
GNU date takes a bunch of extra useful params like that eg http://www.cyberciti.biz/tips/linux-...rows-date.html

I see no need to reset your PATH env var and then set vars pointing to the various binaries:
I doubt very much that /bin and /usr/bin is not in your PATH env var. It's much cleaner to leave $PATH as is and define a var for your data path:
and then append $FILE_NAME to this:
It also means you can call sed with:
instead of I think that would make the code cleaner and a lot easier to read.

Also, since the shell uses all uppercase letters for it's env vars, it's always recommended that you use a different naming convention for your script vars. Instead of:
use