When you are working with Apache Iceberg tables, you can remove snapshots that you no longer need and the files (data files, the manifest file, manifest list file, and partition stats files) that are associated only with them. See Expiring Snapshots of Apache Iceberg Tables for more information.Syntax
VACUUM TABLE <table_name>
EXPIRE SNAPSHOTS [older_than <value>] [retain_last <value>]
The path and name of the table to which the snapshots to expire belong.
EXPIRE SNAPSHOTS [ older_than <value> ][ retain_last <value> ]
Options for specifying which snapshots to expire.
older_than: Timestamp that marks the lower boundary of the snapshots to keep. Snapshots with timestamps earlier than the specified timestamp are deleted. The default timestamp is five days before the date and time at which the command is run.
retain_last: Minimum number of snapshots to retain, starting with the current timestamp.
Example 1Remove snapshots older than specific day and time, but retain a minimum of 20 snapshots
VACUUM TABLE 's3.SF_Incidents2016'
EXPIRE SNAPSHOTS older_than '2023-04-20 00:00:00.000' retain_last=20;
The most recent twenty snapshots are retained. For example, suppose that eleven snapshots are created after the date and time represented by the timestamp. Eight snapshots older than the timestamp are retained, so as to retain the minimum amount of twenty.
Example 2Remove snapshots older than 5 days, but retain a minimum of 100 snapshots
VACUUM TABLE 's3.SF_Incidents2016' EXPIRE SNAPSHOTS retain_last=100;
The most recent one hundred snapshots are retained, even among snapshots that are older than five days. After the 100 snapshots, any snapshots older than 5 days are deleted. For example, suppose that eleven snapshots were created in the past five days. When the command is run, the most recent one hundred snapshots are retained, even though the majority of them were created further back in the past than five days ago.
Example 3Remove snapshots older than 5 days, but retain a minimum of 1 snapshot
VACUUM TABLE 's3.SF_Incidents2016' EXPIRE SNAPSHOTS;
The snapshots created within the last five days are retained. For example, suppose that eleven snapshots were created in the last five days. Only those snapshots would be retained.
The VACUUM TABLE command is not supported for tables in Arctic catalogs. Use the VACUUM CATALOG command to remove snapshots and orphaned metadata files for tables in Arctic catalog sources.