CLUSTERING

Function

Performs the clustering operation on Hudi tables. For details, see Hudi Clustering.

Syntax

Performing clustering:
call run_clustering(table=>'[table]', path=>'[path]', predicate=>'[predicate]', order=>'[order]');
Viewing the clustering plan:
call show_clustering(table=>'[table]', path=>'[path]', limit=>[limit]);

Parameter Description

**Table 1** Parameter descriptions
Parameter	Description	Mandatory
table	Name of the table to be queried. The value can be in the database.tablename format.	Either table or path must be set.
path	Path of the table to be queried	Either table or path must be set.
predicate	Predicate to be defined, which is used to filter partitions to be clustered	No
order	Sorting field for clustering	No
limit	Number of query results to display	No

Example

call show_clustering(table => 'hudi_table1');

call run_clustering(table => 'hudi_table1', predicate => '(ts >= 1006L and ts < 1008L) or ts >= 1009L', order => 'ts');

call run_clustering(path => 'obs://bucket/path/hudi_test2', predicate => "dt = '2021-08-28'", order => 'id');

Caveats

Either table or path must exist. Otherwise, the Hudi table to be clustered cannot be determined.
When using the metadata service provided by DLI, this command only supports configuring the table parameter and does not support configuring the path parameter.
To cluster a specified partition, refer to the format predicate => "dt = '2021-08-28'".