Hi folks,
In the case of Range partition I can clearly see and understand the value. ie: if I create a range partition on MSEG based upon Material Doc year (MJAHR) I can have a partition for each year. Let's say I have 10 years and 10 partitions. If I query the year 2006 then only that partition is read and the advantage is obvious and I get much faster response times.
But in contrast, if I were to create a HASH partition it is supposedly an equal distribution. So let's say I created a HASH 10 partition on sales item table VBAP. Since there is no nice key field based on YEAR in this table like there is in MSEG lets say I use Sales Document Number (VBELN) and create HASH 10 VBELN. Now lets say our sales document numbers start at 0000000001 all the way through 9999999999. Now I run a query that pulls all sales documents with number somewhere in the middle such as 5555555555. Couldn't this sales document exist in ANY of the 10 equally distributed partitions? Meaning wouldn't each and every partition need to be searched to find the value 5555555555? This is the part that I'm trying to understand about HASH partition. Is there still some sort of unseen range within the 10 hash buckets? I need to understand the algorithm how these records are distributed because if ALL hash partitions are read when I run my query then I do not see any advantage in this type of partition.
Many thanks to anyone who can help clarify HASH type partition!
-Patrick