diff --git a/Content/change-types/databricks/alter-cluster.html b/Content/change-types/databricks/alter-cluster.html index a8e9651e8..a94a073f6 100644 --- a/Content/change-types/databricks/alter-cluster.html +++ b/Content/change-types/databricks/alter-cluster.html @@ -14,7 +14,6 @@

Uses

Clustered columns can help optimize performance for some database queries. If you have previously created a table with one or more clustered columns, you can modify which columns are clustered using alterCluster. Specify which columns to cluster using clusterBy.

Changing which columns are clustered can be useful if your data changes significantly or if you begin using different filters to query your data. Better clustering can improve the read efficiency of the new queries.

Databricks does not allow you to drop tables containing clustered columns. You can use alterCluster to remove clustering and then drop the table.

-

does not automatically re-cluster table rows you update with alterCluster. To do this, run optimizeTable after altering your table.

For more information, see Use liquid clustering for Delta tables and ALTER TABLE.

Run alterCluster

@@ -49,10 +48,10 @@

clusterBy

Optional.

-

Specifies how to cluster the table.

+

Specifies how to cluster the table. Use this to remove clustering from a column.

clusterBy has the following nested attributes:

columns/column

Optional. @@ -141,6 +140,8 @@

Examples

+

Troubleshooting

+

If you set the clusterBy  none=false, throws this error:

Unexpected error running Liquibase: Error parsing line 13 column 49 of generated.xml: cvc-enumeration-valid: Value 'false' is not facet-valid with respect to enumeration '[true]'. It must be a value from the enumeration.

Related links