The most performance friendly way to do this ?

Question

I have a program thats in PROD and I want to make modifications to it.
The program loads an internal table with a high volume of data and does a bunch of processing on each record. 
I need to modify the code so that when I do the first Loop and I am processing a rectype 1 record,  I now want to consolidate it with other records eg:  
ctry rectype    pid  date  time etc ...
CA    1          2
CA    1          2
For every rectype 1 with the same pid, perform checks and consolidate the records into 1. 
We are processing high volume data.  So I would like to implement the most performance freindly modification.
What is the best way to do the consolidation?
So when I get the first rectype 1 within my first loop, I need to go get all the records with the same pid.  Consolidate them and then delete them so I dont pick them up again.
thx for your input .. J

Answer 1

If I understand this properly... Big if!
1) create a duplicate table
2) sort both tables by PID
3) delete duplicate PIDs from the main table. This will become your consolidated table
4) loop through the new "summary" table and loop through the original "details" table and do your consolidation logic. Use the initial binary read to identify the starting point in the table of the PID to be consolidated. This will avoid a full table scan for each summary record. Modify the summary table with the updated data.
This will be performance "friendly"
li_summary[] = original_table[].
li_details[] = li_summary[].
SORT: li_details BY pid,
      li_summary BY pid.
DELETE ADJACENT DUPLICATES FROM li_summary COMPARING pid.
LOOP AT li_summary ASSIGNING <summary>.
  READ TABLE li_details TRANSPORTING NO FIELDS
      WITH KEY pid = <summary>-pid
      BINARY SEARCH.
  IF sy-subrc = 0.
    LOOP AT li_details assinging <same_pid> FROM sy-tabix.
      IF <same_pid>-pid <> <summary>-pid.
        EXIT.
      ENDIF.
consolidate or whatever you need to do
     <summary>-??? = ?? + ??? etc.
    ENDLOOP.
  ENDIF.
ENDLOOP.
original_table[] = li_summary[].