Shell Programming and Scripting

 awk -F":" 'NR==FNR{++a[$1,$3,$4];next} !(a[$1,$3,$4])' file1.txt file2.txt

123:09-15-2011:abc:123456
123:09-15-2011:abc:234567
123:09-15-2011:abc:345678

123:09152011:abc:123456
123:09152011:abc:234567
123:09152011:abc:124567

$ awk 'BEGIN { FS = OFS = ":" } NR==FNR{ $2 = ""; ++a[$0];next} { second_field = $2; $2 = ""; if ( !(a[$0]) ) { $2 = second_field ; print } }' file1 file2
123:09152011:abc:124567

$ cat script.awk
## Execute this part once. Set input and output field separators to ':'.
BEGIN {
        FS = OFS = ":"
}

## FNR counts lines of each file and NR counts lines of all input files so
## they only will be equal processing first input file.
NR == FNR {
        ## Remove second field and save all the line in an array, so lines will be
        ## saved like:
        ## 123::abc:123456
        ## 123::abc:234567
        ## 123::abc:345678
        ##
        ## We have rid of second field!!
        $2 = ""
        ++a[$0]

        ## Process next line from the beginning.
        next
}

## Run this part of code in every line (of second file).
{ 
        ## Save second field and reset it, like in first file.
        second_field = $2
        $2 = ""

        ## Search for the line in the array, if not exists it means that any field 
        ## (except second one, because was empty) is different, so recover second field and
        ## print the line.
        if ( !a[$0] ) {
                $2 = second_field
                print
        }
}
$ awk -f script.awk file1 file2
123:09152011:abc:124567