QIF Adder (updated to remove duplicate entries)

I do my banking online, and my bank provides QIF files (Quicken files) for download. The QIF files are plain text files with a simple format. Here's a sample:

D08/10/2006
T-5.81
C*
N
PPurchase
M# XXXXXXXXX GIANT FOOD INC #XXX CITY,STATE 08/09/06
^

Each entry ends with a "^". This particular transaction was a withdrawal of $5.81. Now software such as gnucash and grisby can read qif files, but they do not quite work for me. The balance they come up with does not match the balance reported by my bank. Needless to say, when it comes to balancing my checkbook, the last thing I want to deal with is buggy software.

So I decided to write a shell script called qifadder to read a QIF file and add all of the entries to an opening balance. The script expects a file OpenningBalance to exist in the same directory. This file should simply contain a number giving the opening balance of the account.

I run the script like this (the purpose of using tac, cat in reverse, is to have the script use the same transaction order as it appears on my bank's website):

$ ./qifadder fib_01-01-06_to_8-19-06.qif > summary.out; tac summary.out > summary.tac

Here is a listing of qifadder (Updated to remove duplicates):

#!/bin/sh
# Usage: qifadder file.qif
# Simple script to add up transactions in a qif file.

tac $1 > $1.tac
read openbal < OpenningBalance

bal=$(echo "scale=2; $openbal" | bc)

count=0
while read line
do
firstChar=${line:0:1}
case "$firstChar" in
^ ) lineCarrot["$count"]="$line";;
M ) lineM[count]="$line";;
P ) lineP[count]="$line";;
N ) lineN[count]="$line";;
C ) lineC[count]="$line";;
T ) lineT[count]="$line";;
D ) lineD[count]="$line"
# Now all lines have been read in for the transaction.
# Search through previous transactions to check for duplicates.
# If the date has changed, reset count to 0 so that the
# search list does not keep building up.
prevcount=0
foundDup=0
while [ "$prevcount" -lt "$count" ]
do
if [ "${lineD[count]}" = "${lineD[0]}" ]
then
if [ "${lineT[count]}" = "${lineT[prevcount]}" ] &&
[ "${lineC[count]}" = "${lineC[prevcount]}" ] &&
[ "${lineN[count]}" = "${lineN[prevcount]}" ] &&
[ "${lineP[count]}" = "${lineP[prevcount]}" ] &&
[ "${lineM[count]}" = "${lineM[prevcount]}" ] &&
[ "${lineCarrot[count]}" = "${lineCarrot[prevcount]}" ]
then # found duplicate
echo "found duplicate"
foundDup=1
fi
else # restart the list to search for duplicates
lineD[0]="${lineD[count]}"
lineT[0]="${lineT[count]}"
lineC[0]="${lineC[count]}"
lineN[0]="${lineN[count]}"
lineP[0]="${lineP[count]}"
lineM[0]="${lineM[count]}"
lineCarrot[0]="${lineCarrot[count]}"
count=0
fi
let "prevcount = $prevcount + 1"
done

if [ "$foundDup" = 0 ]
then
# line*[count] is not a duplicate, process the line
tmp="${lineT[count]}"
trans="${tmp:1}" # keep the number after "T"
trans=`echo $trans | sed -e 's/,//'` # get rid of commas
bal=$(echo "scale=2; $bal+$trans" | bc)
echo "change: $trans balance: $bal"
fi

let "count = $count + 1"
;;
esac

done <"$1.tac"

echo ""
echo "balance is $bal"