Page tree
Skip to end of metadata
Go to start of metadata


Peter Huang, Rong Rong, Bowen Li, Shuyi Chen


Current state: 1.10 part - Accepted, the other part is still under discussion

Discussion thread



Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).


This proposal aims to support function DDL with the consideration of SQL syntax, language compliance, and advanced external UDF lib registration.

The Flink DDL is initialized and discussed in the design [1] by Shuyi Chen, et. al. As the initial discussion mainly focused on the table, type and view. FLIP-69 [2] extend it with a more detailed discussion of DDL for catalog, database, and function. Original the function DDL was under the scope of FLIP-69.

After some discussion with the community, we found that there are several ongoing efforts, such as FLIP-64 [3], FLIP-65 [4], and FLIP-78 [5]. As they will directly impact the SQL syntax of function DDL, the proposal wants to describe the problem clearly with the consideration of existing works and make sure the design aligns with efforts of API change of temporary objects and type inference for UDF defined by different languages.

Proposed Changes


Before deep into the DDL SQL, we want to discuss the major requirements for defining a function within Flink runtime by related FLIPs:

  • External lib registration: The requirements come from the hive integration that enhances the adoption of Flink batch. HQL supports syntax like:
CREATE FUNCTION addfunc AS 'com.example.hiveserver2.udf.add' USING JAR 'hdfs:///path/to/jar'
  • Language Distinction: Due to bytecode language specifics in Scala, there are some limitations to extract type information from scala function. At the same time, support python UDF in table runtime is another ongoing effort. Thus, the SQL syntax needs to consider supporting multiple languages. For example: Mysql create function syntax support language in this way: 


  • Temporary Function Support: FLIP-57 proposes to distinguish temporary and non-temporary functions for both catalog and system. As a temporary function will be registered only for the current session. It requires a flag from DDL to distinguish the function resolution order.
 CREATE TEMPORARY FUNCTION addfunc AS 'com.example.hiveserver2.udf.add' USING JAR 'hdfs:///path/to/jar'
  • Function Qualifier: Functions identifiers resolution consider object scopes whether in particular catalog, database or just current catalog and database. Thus, all of the function DDL needs to support 3-part path.

CREATE FUNCTION catalog1.addfunc AS 'com.example.hiveserver2.udf.add' LANGUAGE JAVA

Function DDL Syntax

We propose the following as the function DDL syntax:

Create Function Statement


Drop Function Statement

DROP [TEMPORARY|TEMPORARY SYSTEM] FUNCTION [IF EXISTS] [catalog_name.][db_name.] function_name;

Alter Function Statement

ALTER [TEMPORARY| TEMPORARY SYSTEM] FUNCTION [IF EXISTS] [catalog_name.][db_name.] function_name RENAME TO new_name;

Show Function Statement

SHOW FUNCTION  [catalog_name.][db_name]

We currently only support java/scala/python. Both java and scala run in JVM. Technically, JVM and python are enough to distinguish two runtimes in Flink. But JVM and python are conceptually in different domains as JVM is runtime and python is language. Thus, we distinguished JAVA and SCALA in DDL syntax.

Use Cases

We want to use the function syntax to support all potential use cases. Below we list some obvious use cases that can be achieved.

Load UDF from Classpath


DROP FUNCTION catalog1.db1.geofence

In this case, the assumption is that the UDF classes are already in the Classpath. Thus, we just need to get the class object by reflection, determine whether it is UDF, UDAF or UDTF,  and register it to TableEnvironment.

Load UDF from a remote resource

In this case, the user can use a class that is not the local Classpath. In the example above, the function NestedOutput is contained in a jar that is released to Artifactory.

Using this type of model, we can split the user level logic from the platform. Each team can write and own its own UDF library. A Flink platform is just responsible to load it into Classpath and use it. We will discuss how to achieve it in the later section. Basically, the resource URL will be added as a user library in the execution environment. It will be added into a job graph, and ship to the storage layer, such as HDFS before job submission.

Load python UDF from a remote resource

CREATE FUNCTION  catalog1.db1.func3 AS ''  LANGUAGE 'PYTHON' USING 'http://external.resources/'

New or Changed Public Interfaces

The first change needed is to add more functions in CatalogFunction interface.

public interface CatalogFunction {

  String getClassName();

  Enum getLanguage();  // TODO

  Map<String, String> getProperties();

  CatalogFunction copy();

  Optional<List<String>> getResourcePaths();  // TODO

  Optional<String> getDescription();

  Optional<String> getDetailedDescription();


The second change is to register user Jar: In order to support loading external libraries and create UDFs from external libraries, we need to add a function in ExecutionEnvironment to register external libraries.


 * Register a jar file to load in the Flink job dynamically. The jar file will be added into job graph before job   

 * submission. During runtime, the jar file is loaded into user code class loader automatically.


 * @param jarFile The path of the jar file (e.g., “file://a”, “hdfs://b”, “http://c”)


Public void registerUserJarFile(String jarFile) {

    Path path = new Path(jarFile);



Before the job submission, register user jars will be added into StreamGraph, and then be added into JobGraph in the JobGraphGenerator. 

Resource Isolation

To consider the isolation of class loading of different session, we can add a new interface in {Stream}ExecutionEnvironment. Such as:

Public void registerUserJarFiles(String classloaderName, String... jarFiles) {

  // ...


The interface register a set of Jar files with a specific Classloader environment key: classloaderName. Internally, it uses similar path as registerCachedFile(), which distributes the Jar files to runtime using Flink’s Blob Server. 

Also, add a new interface in RuntimeContext to create and cache a custom userCodeClassLoader using the Jar file set registered under name

Public ClassLoader getClassLoaderByName(String classloaderName) {

  // ...


During code generation of the UDF function call, it will load the set of Jar files that are associated with the library into a custom classLoader, and invoke the function reflectively.

Also, inside RuntimeContext implementation, we will keep a cached of all loaded custom classLoader so we wont load the same library multiple times. 

Implementation plan

Note: For different language support, the implementation will be different, here we just discuss the API needed for java/scala libs.

From an implementation perspective, we want to provide function syntax align with multiple language support, but limit the scope only in java and scala. The python udd related support will be discussed and implemented in the scope of FLIP-78. The concrete action items include

Flink 1.10 Release

In the Flink 1.10 release, we will focus on the basic function DDL syntax as below:

Create Function Statement


Drop Function Statement

DROP [TEMPORARY|TEMPORARY SYSTEM] FUNCTION [IF EXISTS] [catalog_name.][db_name.] function_name;

Alter Function Statement


Show Function Statement

SHOW FUNCTIONS  [catalog_name.][db_name]

Basically, we will delivery create function and drop function that is already included in the classpath. For loading function from remote resources, it will be work after the Flink 1.10 release. The sub-tasks include.

  1. Add function related syntax in Flink SQL parser.
  2. Define SqlCreateFunction and SqlDropFunction in flink-sql-parser module
  3. Add changes to CatalogFunction
  4. Bridge DDL to register the function into Table environment and catalogs

After Flink 1.10

  1. Support UDF loading from an external resource for java
  2. Add scala function related support

As FLIP-65 New type inference for Table API UDFs is a blocker for adding scala function into TableEnvImpl. 1), 2) and 3) will only support language java. 4) is for adding a function

into the table environment with remote resources.  Once the FLIP-65 is done, we can continue the work of supporting language Scala, and corresponding function registration into TableEnvImpl.

Migration Plan and Compatibility

It is a new feature for Flink DDL, there is no migration needed.

Rejected Alternatives

The function DDL syntax considered about existing requirements. No rejected alternatives yet.


[1] Flink SQL DDL Design  

[2] FLIP-69 Flink SQL DDL Enhancement

[3] FLIP-64 Support for Temporary Objects in Table module

[4] FLIP-65 New type inference for Table API UDFs

[5] FLIP-78 Flink Python UDF Environment and Dependency Management

  • No labels