These restrictions apply:
You cannot use ALC to compress columns that have a data type of structured UDT.
The TD_LZ_COMPRESS and TD_LZ_DECOMPRESS system functions compress all large UDTs including
UDT-based system types such as Geospatial, XML, and JSON. However, if you write your
own compression functions, the following restrictions apply:
Custom compression functions cannot be used to compress UDT-based system types (except
for ARRAY and Period types).
Custom compression functions cannot be used to compress distinct UDTs that are based
on UDT-based system types (except for ARRAY and Period types).
You cannot write your own compression functions to perform algorithmic compression
on JSON type columns. However, Teradata provides the JSON_COMPRESS and JSON_DECOMPRESS
functions that you can use to perform ALC on JSON type columns.
You cannot use ALC to compress temporal columns:
A column defined as SYSTEM_TIME, VALIDTIME, or TRANSACTIONTIME.
The DateTime columns that define the beginning and ending bounds of a temporal derived
period column (SYSTEM_TIME, VALIDTIME, or TRANSACTIONTIME).
You can use ALC to compress Period data types in columns that are nontemporal; however,
you cannot use ALC to compress derived period columns.
For details about temporal tables, see Temporal Table Support and ANSI Temporal Table Support.
You cannot specify multi-value or algorithmic compression for a row-level security
You can apply algorithmic compression to referential integrity columns.
Depending on the implementation, algorithmic compression can be either physical or
logical, though most implementations use physical data compression. Algorithmic compression
can be either lossy or lossless, depending on the algorithm used.