Hello Lars,
Appreciate a lot of your quick reply, cannot believe I got your reply!!! You are the super star on SCN of HANA topic, it's my honor to discuss with you!!!! Thanks a lot!
The concerns I had is about the confusion on table size in memory v.s table size on disk. Let take this customer's case for example. After customer migrate BW on DB2 to BWon HANA, a few BW table has big size on disk. For instance, table /BIC/AZI_BLS0100 , on disk it is 109GB, but in memory (full load, 6 partition on 6 nodes) , total memory used is around 42GB.
According to the compression ratio, that's a very good point, I attached compression ration for this table on each columns, looks some columns good some ones are not great..
Because this table on disk was wrapped into data volume size, customer now facing on several /BIC/xxxxx tables had the same situation, so that the data volume is huge that make data backup size would be horrible. The other thing I should mentioned here is this customer do the same system migration 2 months ago, same BW system to same HANA system, HANA revision is 85.02 now, this time the data size on BW even smaller than 2 months ago. However, at that time, data volume was 5T on disk, this time was increased to 9T. That's very interesting but concerns customer, they are worried the sizing is not enough on disk, they are also concerned that the data backup would be uncontrolled huge to maintain after go live.
So I would like to confirm that to my understanding, table on disk not doing compression, only if it load to memory it will be compressed, am I right? and if you can share any good ideas that what caused size on disk of those /BIC/XXXX table is much bigger than in memory, that will be awesome. BTW, I try to compare to the other customer's BWon HANA system, in the other customer's system, /BIC/XXXX table size on disk is almost the same size in memory.
Thank you again, Lars!!! Appreciate all your effort on any comments.
-Tian =)