Only key fields of table-cluster in where-condition for better performance?

Question

Hi,
Say my cluster table is having following key fields: KEY1, KEY2, KEY3, KEY4.
And, my table-cluster is having following key fields: KEY1, KEY2, PAGENO .. [PAGENO is always there in a table-cluster].
Now, I have two versions of my (Open SQL) SELECT-WHERE clause:
a) SELECT *
    WHERE KEY1 EQ LV_KEY1
         AND KEY2 EQ LV_KEY2
         AND KEY3 EQ LV_KEY3.
b) SELECT *
    WHERE KEY1 EQ LV_KEY1
         AND KEY2 EQ LV_KEY2.
Please guide me whether version (a) or version (b) will be executed faster? In case of version (b), I will have to filter data using ABAP code corresponding to KEY3.
Regards,
Kamlesh Dhera H. R.

Answer 1

I think most experienced ABAPrs would tell you to try both, compare runtimes, and use the one that's more efficient.  SAP provides "get run time field <fieldname of type i>.  Get run time at start, get run time at end, and subtract start from end to get the time in microseconds (1 second / 1,000,000 ).
Run several times to minimize effect of buffering and system load activity, and you'll see which method produces lower run time.