Partitioning a Table or Join Index That Does Not Have a Primary Index
Besides partitioning tables or join indexes on their rows, you can also partition them on their columns. Column partitioning enables Teradata Columnar. Column partitioning is only allowed when a table or join index has no primary index (see “NoPI Tables, Column‑Partitioned Tables, and Column-Partitioned Join Indexes” on page 280).
Column partitioning is a physical database design choice that is not suitable for all workloads. For example, column partitioning is usually not suitable for workloads that often select both a significant number of rows and project many columns from a table. However, column partitioning might be suitable if a request selects a significant number of rows, but projects only a few columns, or conversely, if a request projects many columns, but only selects a small number of rows.
Column partitioning is especially suitable for the case where both a small number of rows are selected and only a few columns are projected. See “Column‑Partitioned Tables and Join Indexes” on page 285 for more information about how table and join indexes can be partitioned on their columns and how column‑partitioned tables and join indexes can be applied optimally.
Note: A column‑partitioned table or join index can also have 1 or more row partitioning levels.