15.00 - Storage and Other Overhead Considerations for Partitioning - Teradata Database

Teradata Database Design

prodname
Teradata Database
vrm_release
15.00
category
User Guide
featnum
B035-1094-015K

Storage and Other Overhead Considerations for Partitioning

Implications for Storage of Partitioning

The storage implications of partitioned tables and join indexes are minimal.

  • Each row in a partitioned table or join index has either 2 or 8 additional bytes in its row header to store its internal partition number.
  • An empty partition occupies no space on disk, and a populated partition occupies only 2, 4, or 8 bytes more per row than the same rows would if they belonged to a nonpartitioned table because a partition is just an ordering of the rows on an AMP.
  • There is no additional overhead accrued by inserting the first row into a partition.
  • Another way to say this is that there is no action required to create a partition. Rows are just inserted in their partition/row hash/uniqueness order.

  • There is no additional overhead accrued by deleting the last row of a partition.
  • In other words, there is no additional action required to drop a partition. The last row of a partition is just deleted in the same way as any other row.