Renaming first column to distinguish between startsites script - linux

Question

Im looking for a way to convert the file as below called INPUT to OUTPUT. The file INPUT consists of columns consisting the unique ID, ID and the value. I would like to convert the ID to separated IDs based on the value as distinction. I tried some basic commands but could not manage to make it work for the main input file which is 20,000 rows and has 15,000 IDs.
Does anyone has some nice ideas/suggestions how to handle this problem?
INPUT OUTPUT
unique ID VALUE unique ID VALUE
A1 GENEA 10 -> A1 GENEAp1 10
A2 GENEA 5 -> A2 GENEAp2 5
A3 GENEA 2 -> A3 GENEAp3 2
A4 GENEB 4 -> A4 GENEBp4 4
A5 GENEB 5 -> A5 GENEBp3 5
A6 GENEB 8 -> A6 GENEBp2 8
A7 GENEB 70 -> A7 GENEBp1 70
A8 GENEC 5 -> A8 GENECp1 5
A9 GENED 50 -> A9 GENEDp2 50
A10 GENED 10 -> A10 GENEDp3 10
Preferably the numbering of p based on the value. With p1 with the highest value, p2 second highest etc.

Answer 1

Here's a crazy one-liner that does it:
head -1 file; tail -n+2 file| nl| sort -nrk4| awk '{ ++m[$3]; print($1" "$2" "$3"p"m[$3]" "$4); }'| sort -n| cut -d' ' -f2-4| column -to' ';
Output:
unique ID VALUE
A1 GENEAp1 10
A2 GENEAp2 5
A3 GENEAp3 2
A4 GENEBp4 4
A5 GENEBp3 5
A6 GENEBp2 8
A7 GENEBp1 70
A8 GENECp1 5
A9 GENEDp1 50
A10 GENEDp2 10
It involves sorting the file by the VALUE column, and then processing it sequentially in awk, counting occurrences of each distinct ID in an associative array, so you can build up the p# count.
Additional notes:
I printed the header line (head -1) separately from the data lines (tail -n+2) so the main processing pipeline would only apply to the data lines.
I added a call to nl before the initial sort to capture the original line order in a new leading numbering column, and then sorted by that column afterward (and then cut out that numbering column) to return to the original order.
I added column -to' ' at the end to align the data lines, don't know if you want/need that. If you want to align the header line with the data lines, you can surround the head statement and main pipeline with a braced block and move the column -to' ' filter outside the braced block to align the whole thing.