PTI tables are optimized to store and retrieve time series data. Time series data typically comes from monitoring sensors that regularly record various data, such as climate or ocean conditions, internal operating status of devices and vehicles, locations of GPS tracking sensors, stock prices, and so forth.
Time series tables are similar to standard, non-PTI Vantage tables, but include a primary time index (PTI) rather than a primary index (PI). Like a PI, a PTI can speed access to frequently queried data, determines how the table row data is distributed among the system AMPs (storage), and determines how the rows are ordered within the table. A properly chosen PTI can help ensure an efficient distribution of data that takes best advantage of the parallelism of Vantage.
Restrictions on PTI Tables
- PTI tables cannot be created as standard primary index (PI) or primary AMP index (PA) tables, and cannot be created as NO PRIMARY INDEX (NoPI) tables..
- PTI tables cannot have join or hash indexes.
- PTI tables cannot include row or column partitioning.
- PTI tables cannot also be temporal tables.
- PTI tables cannot also be queue tables.
- PTI tables cannot use load isolation.
For more information about creating standard tables, see Teradata Vantage™ - SQL Data Definition Language Syntax and Examples, B035-1144.