MaxCompute offers a variety of built-in functions to meet your business requirements. If the built-in functions do not meet your business requirements, you can write code to create user-defined functions (UDFs). This topic describes the types, scenarios, development process, and usage notes of UDFs that are supported in MaxCompute.
In the broad sense, UDFs include user-defined scalar functions, user-defined aggregation functions (UDAFs), and user-defined table-valued functions (UDTFs). In the narrow sense, UDFs refer only to user-defined scalar functions. The following table describes the types of MaxCompute UDFs.
|UDF||A one-to-one mapping is established between the input and output data of a UDF. Each time a UDF reads a row of data, it returns an output value.|
|UDTF||A one-to-many mapping is established between the input and output data of a UDTF. Each time a UDTF reads a row of data, it returns multiple values, which are considered a table.|
|UDAF||A many-to-one mapping is established between the input and output data of a UDAF. Multiple input records are aggregated to generate one output value.|
In addition to the preceding UDFs, MaxCompute offers the following UDFs for special scenarios.
|Code-embedded UDFs||If you want to simplify the development process of MaxCompute UDFs and view the code logic, you can embed Java or Python code into SQL scripts.|
|SQL functions||If your code contains duplicate code, you can use SQL UDFs to improve the code reuse rate and simplify the development process.|
|Open source geospatial UDFs||You can use Hive geospatial functions to analyze spatial data in MaxCompute.|
You cannot access the Internet by using UDFs. If you want to access the Internet by using UDFs, fill in the network connection application form based on your business requirements and submit the application. After the application is approved, the MaxCompute technical support team will contact you and help you establish network connections. For more information about how to fill in the network connection application form, see Network connection process.
- UDFs cannot compete with built-in functions in performance. We recommend that you preferentially use built-in functions to implement your business logic.
- If you use a UDF in SQL statements, the memory usage of a computing job may exceed
the default allocated memory size if a large amount of data is computed and data skew
occurs. In this case, you can run the
set odps.sql.udf.joiner.jvm.memory=xxxx;command at the session level to resolve the issue. For more information about the MaxCompute UDF FAQ, see FAQ about MaxCompute UDFs.
- If the name of a UDF is the same as that of a built-in function, the UDF is preferentially
called. For example, if UDF CONCAT and built-in function CONCAT both exist in MaxCompute,
the system automatically calls UDF CONCAT instead of the built-in function CONCAT.
If you want to call the built-in function, you must add the symbol
::before the built-in function, for example,
select ::concat('ab', 'c');.
This section describes how to develop a UDF, UDTF, or UDAF. The development process of Code-embedded UDFs, SQL functions, and Open source geospatial UDFs is different. For more information, see the related documentation.
- The following figure demonstrates how to write a MaxCompute UDF in Java.
No. Required Description Platform References 1 No Before you can use Maven to write code, you must add the related SDK dependencies to the POM file. This ensures that the code can be compiled. The following SDK dependency shows an example:
You can search for
<dependency> <groupId>com.aliyun.odps</groupId> <artifactId>odps-sdk-udf</artifactId> <version>0.29.10-public</version> </dependency>
odps-sdk-udffrom Maven repositories to obtain the version of the SDK dependency.
IntelliJ IDEA (Maven) None 2 Yes Write a UDF based on your business requirements. IntelliJ IDEA (Maven) and MaxCompute Studio Develop a UDF in Java 3 Yes Debug the UDF by running it on your on-premises machine or by performing unit testing to check whether the result meets expectations. 4 Yes Debug the UDF code to ensure that the code is packaged into a JAR file after it is successfully run on your on-premises machine. 5 Yes Upload the JAR file as a resource to your MaxCompute project. MaxCompute client, MaxCompute Studio, and DataWorks
- MaxCompute client
- MaxCompute Studio
6 Yes Create a UDF based on the JAR file that you uploaded. 7 No Call the UDF in the query data code. None
- The following figure demonstrates how to write a MaxCompute UDF in Python.
No. Required Description Platform References 1 Yes Write a UDF based on your business requirements. MaxCompute Studio Develop a Python UDF 2 Yes Debug the UDF by running it on your on-premises machine or by performing unit testing to check whether the result meets expectations. 3 Yes Upload Python files or required resources, such as file resources, table resources, and third-party packages, to a MaxCompute project. MaxCompute client, MaxCompute Studio, and DataWorks
- MaxCompute client
- MaxCompute Studio
4 Yes Create a UDF based on the uploaded Python files or required resources. 5 No Call the UDF in the query data code. None
- Use a UDF in a MaxCompute project: The method is similar to that of using built-in functions.
- Use a UDF across projects: Use a UDF of Project B in Project A. The following statement
shows an example:
select B:udf_in_other_project(arg0, arg1) as res from table_t;. For more information about resource sharing across projects, see Package-based resource sharing across projects.
The following table describes the SDKs provided by MaxCompute. For more information about the packages included in each SDK and the classes in the packages, see MaxCompute SDK.
|odps-sdk-core||Provides classes for managing basic resources of MaxCompute.|
|odps-sdk-commons||Provides common Utils for Java.|
|odps-sdk-mapred||Provides the MapReduce API.|
|odps-sdk-graph||Provides the Graph API.|