A data block is the unit of physical I/O when Analytics Database writes data to disk. A data block contains one or more data rows from the same table. The system stores data blocks in segments, which are grouped in cylinders.
For Block File System tables, you can set data block sizes for tables using SQL. If there is a conflict between the data block size value set for the system and for the table, the table value overrides the global value.
For Object File System tables, data block sizes have a default value that you cannot change.
Data Block Merging
As part of full-table modify operations, where the system sequentially scans the table as part of the operation, Vantage can merge small DBs into larger ones as the modifications are performed. This can make the initial modifications to the table more expensive, but makes future full table modifications and queries cheaper by reducing the number of DBs in the table, which reduces disk I/Os.