At ~1:00 you state that, “Delta is read-optimized, and Main is write-optimized”. You also state that, “Delta is a row-store table and Main is a column-store table”. Your diagram shows, and I agree that, Delta is write-optimized and Main is read-optimized. I believe you should be saying that “Delta is write-optimized, and Main is read-optimized”. I understand that Delta is still used for read operations, but if it is row-based it would be optimized for write operations. Do you agree?
In summary, the process involves the separation of the data into two parts: a main part (Main) which contains most of the data organised in columns, and a delta part (Delta) which records recent changes in the data in the form of delta records organised in rows. Read queries are performed in both parts, while insert, update and delete operations are performed only in Delta. When Delta becomes full, the merge process is activated, where data from Main and Delta are merged and transferred to a new unit (Main2 and Delta2). During this process, continuity of data availability for read queries and write operations is ensured by using the old and new drives. Once the data merge and transfer is complete, the old drives are removed and the new drives are converted to the new main and delta configuration. This merge and transfer cycle is repeated periodically as needed to maintain optimal performance and efficient data management in the SAP HANA database.
At ~1:00 you state that, “Delta is read-optimized, and Main is write-optimized”. You also state that, “Delta is a row-store table and Main is a column-store table”. Your diagram shows, and I agree that, Delta is write-optimized and Main is read-optimized. I believe you should be saying that “Delta is write-optimized, and Main is read-optimized”. I understand that Delta is still used for read operations, but if it is row-based it would be optimized for write operations. Do you agree?
In summary, the process involves the separation of the data into two parts: a main part (Main) which contains most of the data organised in columns, and a delta part (Delta) which records recent changes in the data in the form of delta records organised in rows. Read queries are performed in both parts, while insert, update and delete operations are performed only in Delta.
When Delta becomes full, the merge process is activated, where data from Main and Delta are merged and transferred to a new unit (Main2 and Delta2). During this process, continuity of data availability for read queries and write operations is ensured by using the old and new drives. Once the data merge and transfer is complete, the old drives are removed and the new drives are converted to the new main and delta configuration.
This merge and transfer cycle is repeated periodically as needed to maintain optimal performance and efficient data management in the SAP HANA database.