User GuidesAPI ReferenceRelease Notes
Doc HomeHelp CenterLog In


DROP removes one or more attributes. Dropped unified attributes appear as null columns when you preview.

Using DROP can be more convenient to use than SELECT. For example, to increase performance, use DROP to remove unused columns after running JOIN statements. If you remove an attribute in the unified dataset with DROP, it is then populated with nulls.


DROP <col1>, <col2>, ...
This statement removes all of the specified columns from the input dataset.

DROP behaves in the following ways:

  • It removes a column only if the name you specified exists in the dataset.
  • It does not change the ordering of remaining columns.
  • If a column to be dropped duplicates another column, a “<col> is already being dropped” message appears and removes both columns.
  • DROP is not a reserved word. You can use it as the name of an attribute or a dataset.

important Important: Attempting to DROP the tamr_id has no effect. If you explicitly DROP the required origin_source_name or origin_entity_id column, you must manually recreate and repopulate it. See Managing Primary Keys.


DROP column_one;
DROP column_two, column_three, column_four;