Variable-length columns of Disk Data tables take up a fixed amount of space. For each row, this is equal to the space required to store the largest possible value for that column.
For general information about calculating these values, see Data Type Storage Requirements.
You can obtain an estimate the amount of space available in data files and undo log files by querying the
INFORMATION_SCHEMA.FILEStable. For more information and examples, see The INFORMATION_SCHEMA FILES Table.Note
OPTIMIZE TABLEstatement does not have any effect on Disk Data tables.
Each row in a Disk Data table uses 8 bytes in memory to point to the data stored on disk. This means that, in some cases, converting an in-memory column to the disk-based format can actually result in greater memory usage. For example, converting a
CHAR(4)column from memory-based to disk-based format increases the amount of
DataMemoryused per row from 4 to 8 bytes.
Starting the cluster with the
option does not remove Disk Data files.
You must remove these manually prior to performing an initial
restart of the cluster.
Performance of Disk Data tables can be improved by minimizing
the number of disk seeks by making sure that
of sufficient size. You can query the
diskpagebuffer table to help
determine whether the value for this parameter needs to be