delete two patterns and remove one pattern

Shell Programming and Scripting
1

Friends,

I have .txt file with following format.
START
ABC|Prashant1|Patel1
ABC|Prashant2|Patel2
ABC|Prashant1|Patel1
ABC|Prashant2|Patel2
END

I would like to do:
1) Delete line with START
2) Delete line with END
3) Remove ABC|
4) Delete duplicate records

The following command works fine which deletes line with START and END
sed -e /^START/d -e /^END/d Filename.txt

How do I incorporate task 3 and 4?
NOTE: The file will have more than 500,000 thousand rows.

Thanks in advance for suggestion,
Prashant

2 
nawk -F'|'  'NF==3{print $2,$3}' patel
3

This code is not doing uniq + removes pipe delimiter

use this

grep "|" test | sed 's/^ABC|//g' | sort -u
4

Tostay2003:
What if the data doesnt start with ABC your logic fails

use this
nawk -F'|' 'NF==3{print $2,$3}' patel | sort -u

5

I assumed from the authors description that the first field remains common in the file

Use this if you didnt mean that the first field woudl be same.

grep "|" test | cut -d'|' -f2,3 | sort -u

or with slight amendment to code written by zenith i.e. by adding OFS

6 
nawk -F'|' 'NF==3 && !a[$2,$3]++ {print $2,$3}' patel
7

Thank you all for your reply.

nawk -F'|' 'NF==3 && !a[$2,$3]++ {print $2,$3}' patel

I had made small change in print statment because it was not printing | symbol.

nawk -F'|' 'NF==3 && !a[$2,$3]++ {print $2,"|",$3}' patel

However, it prints the space after and before | symbol.

Thanks,
Prashant

8 
nawk -F'|' 'NF==3 && !a[$2,$3]++ {print $2, $3}' OFS='|' patel