This page answers common questions about getting started with MaxCompute, covering required skills, projects, access credentials, storage, and error messages.
-
How to understand open source and cloud-native big data technologies and products?
-
If I disable my AccessKey pair and create a new one, are my auto-triggered nodes affected?
-
Does MaxCompute compress table data automatically? Can I choose the storage format?
Getting started
What skills do I need to use MaxCompute?
MaxCompute supports multiple computing models and data tunnels, so proficiency in SQL, Python, or Java is sufficient to run data analysis workloads.
How to understand open source and cloud-native big data technologies and products?
For more information, see Big Data in Practice: From Open Source to Cloud-Native.
Does MaxCompute support business data monitoring?
MaxCompute supports data monitoring only through the Data Quality feature in DataWorks. Configure data monitoring rules in DataWorks to track your business data.
MaxCompute cannot monitor field-level changes in external data sources.
Projects and access control
What is a MaxCompute project?
A project is the basic organizational unit in MaxCompute — similar to a database or schema in a traditional database system. Each project isolates users and controls access, and can contain tables, resources, functions, and instances. You can have permissions across multiple projects, and cross-project object access is available after security authorization.
How do I get an AccessKey pair?
Go to the AccessKey Pair page to create or view your credentials.
If I disable my AccessKey pair and create a new one, are my auto-triggered nodes affected?
Yes. Disabling or deleting an AccessKey pair causes all nodes in your DataWorks workspace to fail. Auto-triggered nodes created with the previous AccessKey pair stop running.
Proceed with caution when disabling or deleting an AccessKey pair.
Storage and tables
Does MaxCompute compress table data automatically? Can I choose the storage format?
MaxCompute automatically compresses all table data at approximately 6.5:1. The default storage format is AliORC. Customizing the storage format or specifying which columns to compress is not supported.
What table types does MaxCompute support?
MaxCompute supports two table types: internal tables and foreign tables. Foreign table support was introduced in MaxCompute V2.0.
| Internal table | Foreign table | |
|---|---|---|
| Data location | Stored in MaxCompute | Stored in OSS or Tablestore (OTS) |
| What MaxCompute stores | All data | Metadata only |
| Supported column types | All MaxCompute data types | Not specified |
| Typical use case | General-purpose structured data | Unstructured data — video, audio, genetic, meteorological, or geographic data in OSS or OTS |
Development resources
What resources do UDFs and MapReduce jobs require?
-
UDF: After you write a UDF, compile it into a JAR package and upload the package to MaxCompute as a resource. When you run the UDF, MaxCompute automatically downloads the resource package and executes your code.
-
MapReduce: After you write a MapReduce program, compile it into a JAR package and upload the package to MaxCompute as a resource. When you run a MapReduce job, the framework automatically downloads the resource package and executes your code.
You can also upload text files and MaxCompute tables as resources, which UDFs and MapReduce programs can read at runtime.
Error messages
How do I read MaxCompute error messages?
All MaxCompute error messages follow this format:
Error code: General description - Context-related descriptionThe three parts map to the format as follows:
| Part | What it tells you |
|---|---|
| Error code | The category of issue |
| General description | The name of the error type |
| Context-related description | The specific cause in your job |
Error message formats vary slightly across SQL, MapReduce, and tunnel jobs. For a full list of error codes, see Error code overview.