Syntax¶

DROPFUNCTION[IFEXISTS]<name>([<arg_data_type>,...])

Parameters¶

name

Specifies the identifier for the UDF to drop. If the identifier contains spaces or special characters, the entire string must beenclosed in double quotes. Identifiers enclosed in double quotes are also case-sensitive.

arg_data_type[,...]

Specifies the data type of the argument(s), if any, for the UDF. The argument types are necessary because UDFs support nameoverloading (i.e. two UDFs in the same schema can have the same name) and the argument types are used to identify the UDF youwish to drop.

Usage notes¶

All Languages

• Dropped functions can’t be recovered; they must be recreated.

• When the IF EXISTS clause is specified and the target object doesn’t exist, the command completes successfullywithout returning an error.

Java, Python, and Scala

• For UDFs that store code in a file (such as a .jar file or .py file) in a stage, theDROPFUNCTION command does not removethe file. Different UDFs can use different functions/methods in the same file, so the file should not be removed while any UDFrefers to it. Snowflake does not store a count of the number of references to each staged file and does not remove that stagedfile when there are no remaining references.

Examples¶

This demonstrates the DROP FUNCTION command:

DROPFUNCTIONmultiply(number,number);--------------------------------+status             |--------------------------------+MULTIPLY successfully dropped. |--------------------------------+

Copy