In Hive, partition information appears as directories. In Iceberg, partition information is recorded in the final underlying data files, making Iceberg's partitioning more flexible and allowing partition policies to evolve with data volume changes. In Data Lake Compute (DLC), you can create Iceberg tables to use features like hidden partitions.
Note:
Native tables default to Iceberg tables. For external tables, you can choose either Hive or Iceberg tables depending on the file format. For detailed syntax, see the document CREATE TABLE. With hidden partitions, you do not need to specify partition information separately when inserting and querying data, unlike in Hive.
Iceberg partition policies support the following functions. The corresponding partition conversion policies for different fields are shown in the table below:
|
identity | Any | Source Type |
bucket | int, long, decimal, date, time, timestamp, timestamptz, string, uuid, fixed, binary | int |
truncate | int, long, decimal, string | Source Type |
year | date, timestamp, timestamptz | int |
month | date, timestamp, timestamptz | int |
day | date, timestamp, timestamptz | int |
hour | timestamp, timestamptz | int |