on 03-02-2015 6:25 AM
Alright, this seems to be pretty straight forward.
SAP HANA handles new or changed data in a two step approach.
First it saves all data into a insert-only data structure called "Delta Storage" (delta store).
This Delta Storage can handle mass inserts or data changes very well, but at the expense of storage space, as no compression of the data is performed.
The actual data compression is then performed in the second step, when the data gets transferred into the "Main Storage". This in turn is a highly compressed but unchangeable data structure. Once built up, the Main Storage stays as it is.
The process that transfers the data from the Delta to the Main Storage is the "Delta Merge". In short it takes the existing Main Storage and the Delta Storage as input and builds a new combined Main Storage from those.
This new Main Storage is then highly compressed, too.
As you see, it boils down to "No delta merge -> no compression -> high memory usage".
All that is of course described at length in the SAP HANA documentation and in 's and my book ().
- Lars
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello Amir,
Can you try to execute a delta merge on the table by right clicking on the table in HANA Studio and choosing the option 'Perform Delta Merge...'., and then check the usage?
Regards,
Ning Tong
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
101 | |
13 | |
13 | |
11 | |
11 | |
7 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.