Here Key is First 4 letters/length: As per understanding when use SUM FIELDS=NONE it will delete second instance of duplicate record but its not true always ppl says since there few productions scenarios have been failed it seems.
hence i jsut wanted to know whats your view on this ? could you please help me to get better understanding.
Look here for more alternative along with whats being said plus you can add sequence number and restart for each unique combination and use include to get which of the duplicate you want in OUTFIL, SELECT operator
If you are SORTing the data anyway (you need to with what you have shown) you can make the results consistent (a "stable sort" is the technical term for preserving the input order in the case of duplicate keys) by using OPTION EQUALS or specifying EQUAL on the SORT command, as enrico has indicated. No need to do anything else.
With SUM FIELDS=NONE the first record of which there are duplicate keys is retained. For a stable sort, that first record will always be the first record of that key from the input file.
If you specify EQUALS, there is some overhead, because a sequence number (a big one) is always made the most minor part of the sort key. So only use EQUALS when you need it.
"stable sort" doesn't imply anything "unstable", it just means the output is guaranteed reproducible. At a cost.
In a sort which is not stable, the most convenient record at the time will be the one which will be output in the case of duplicate keys.