on 01-29-2015 1:18 PM
Hello,
I read that the compression is performed and optimized automatically in HANA during the delta merge process.
What confuses me is the documentation of the table "TABLE_COLUMNS" (TABLE_COLUMNS - SAP HANA SQL and System Views Reference - SAP Library). I found following text for column COMPRESSION_TYPE = "Type of compression: 'DEFAULT', 'PREFIXED', 'SPARSE', 'CLUSTERED', 'INDIRECT', 'RLE'. Columns in TABLE_COLUMNS show the default value which can be specified during table creation. They can also be updated using an 'ALTER TABLE' statement.”
This text indicates that I can select what type of compression is used manually, which does not fit to the self-managed optimization approach. I had a look in our SAP HANA and I could only find the values “DEFAULT” and “NONE” (which is not even listed here).
Is there a difference between “NONE” and “DEFAULT”? How is it set? (Probably it is just a piece of wrong documentation)
Thanks
Patrik
I assume that NONE is related to row store table columns (where no compression is used) while DEFAULT is related to column store table columns. Obviously the documentation is wrong. I am sorry for that, I will address this SAP internally.
The correct column store compression types can be found in view M_CS_COLUMNS. See SAP Note 2112604 for more information related to SAP HANA compression.
Theoretically there are commands to force a specific compression type for a specific column, but usually the SAP HANA automatisms are doing a good job and so no manual adjustments of compression types are required.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
NONE is indeed related to all non-column store tables columns, since SAP HANA compresses only column store tables. So NONE would be also present for virtual tables or dynamic tiering tables.
The general TABLES/COLUMNS system views typically contains detail attributes that are only valid for specific table types.
But, sure enough, there's a gap in the documentation on this.
- Lars
User | Count |
---|---|
85 | |
23 | |
11 | |
9 | |
8 | |
5 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.