All Products
Search

This Product

    Platform For AI:Read Table

    Document Center

    Platform For AI:Read Table

    Last Updated:Apr 18, 2026

    The Read Table component reads data from MaxCompute tables. By default, the component reads the table data of the current project. To read data from tables in other projects, you must prefix the table names with the related project names.

    Limitations

    The Read Table component supports only the two-level MaxCompute table structure (project.table) and does not support the three-level project.schema.table structure.

    Note

    Currently, Machine Learning Designer algorithm components support the following data types: DOUBLE, BIGINT, STRING, BOOLEAN, and DATETIME. The Read Table component can read all data types. To prevent errors in downstream algorithm components, you can connect a data conversion component downstream of the Read Table component to convert the data types.

    You can configure the parameters of the Read Table component only in the visual interface of Machine Learning Designer. The following table describes the parameters.

    Tab

    Parameter

    Description

    Select Table

    Table Name

    The name of the MaxCompute table from which you want to read data. If you want to read data from a MaxCompute table in another project, set this parameter in the Project name.Table name format. Example: tianchi_project.weibo_data.

    Important

    If you modify a MaxCompute table's columns, Machine Learning Designer cannot automatically synchronize the changes. You must manually reconfigure the MaxCompute data source.

    Partition

    For partitioned tables, select the Partition check box and use the Parameter field to query data by partition.

    Note

    If an SQL Script component is connected downstream, this partition filter condition has no effect. You must configure the filter condition in the SQL Script component.

    Parameter

    The partition that you want to filter. This parameter is used to filter data of the partitioned table. Separate multiple partitions with commas (,). If you do not specify this parameter, the full table is used. Examples:

    • For a dynamic partitioned table, enter a value in the dt=${yyyyMMdd} format. ${yyyyMMdd} is a system parameter that represents the current date and ${yyyyMMdd-1d} represents the day before the current date.

    • For a standard partitioned table, enter partition conditions. Example: ds=1,ds=2.

    • For a multi-level partitioned table, use a forward slash (/) to separate partition levels. For example: ds=1/name=cat,ds=2/name=dog.

    • Use global variables. Example: ds=${a}/name=${b}/tag=${c}. You can configure a, b, and c on the Global Variables tab of the pipeline.

    Fields Information

    Source Table Columns

    The system automatically displays the table schema after you specify the Table Name.