OPEN SQL performance question

Question

Hi friends,
I'm going to read and process data in an interface coded in ABAP and OPEN SQL. To improve efficiency and reliability I'm processing the data in packets of a fixed size of rows - reading rows up to a predetermined numer into an internal table which then is processed and then finaly written back to database followed by "commit work". Then the process will continue with reading the next fixed number of rows, process them, and so on ...
The general question is, which is the most efficient way to implement this scenario?
I think of two basic approaches:
1.1) Loop over results from a cursor using FETCH NEXT CURSOR inside a LOOP appending the lines to the internal table.
2.1) Execute SELECT ... INTO TABLE <itab> FROM <table> UP TO <data packet size> ROWS.
My assumtion is that approach 2 would be the more effecient, is that correct?
The processed data will be written back to the database in one single statement:
2.2) INSERT <table> FROM TABLE <itab>
Which I assume is more efficient than doing the same using multiple inserts within a loop?
Regards,
Christian

Answer 1

In native SQL you can also use the packet options.
SELECT  <Fields name>      appending corresponding fields of table <Internal table>
            <b>package size 20000</b>
            FROM <Database table name>
            WHERE <Condition>.
ENDSELECT.
By using this the system will fetch the records from database table in packets [20000 records per package]
Regards
Aman