This topic introduces Python on MaxCompute (PyODPS) and describes the general usage of PyODPS.

Background information

PyODPS is MaxCompute SDK for Python. PyODPS supports the DataFrame framework and basic operations on MaxCompute objects. You can use PyODPS to analyze data in MaxCompute.

PyODPS supports Python 2.6 and later, and Python 3.

For more information about PyODPs, see the following documentation:
If you want to help build the PyODPS ecosystem, you can perform the following operations:
  • Contribute to the writing PyODPS documentation.
  • Develop PyODPS code at GitHub.
  • For technical support, join the DingTalk group by searching for the DingTalk group ID 11701793.


Before you can use PyODPS, you must initialize a connection to MaxCompute by using your Alibaba Cloud account. To initialize a connection, run the following command:
from odps import ODPS
odps = ODPS('<your_accesskey_id>', '<your_accesskey_secret>', '<your_default_project>', endpoint='<your_end_point>')
  • your_accesskey_id: the AccessKey ID of your Alibaba Cloud account. Make sure that the account has the permissions to manage objects in a MaxCompute project. For more information about the permissions, see Operation permissions. In the upper-right corner of the MaxCompute console, click the profile picture and select AccessKey Management to obtain the AccessKey ID.
  • your_accesskey_secret: the AccessKey secret of the AccessKey ID. In the upper-right corner of the MaxCompute console, click the profile picture and select AccessKey Management to obtain the AccessKey secret.
  • your-default-project: the name of your MaxCompute project. You can log on to the MaxCompute console. In the top navigation bar, select a region and view the name of your MaxCompute project on the Project management tab.
  • your_end_point: the endpoint of the region where your MaxCompute project resides. For more information, see Endpoints.


The following table describes the methods that you can use to perform basic operations on MaxCompute objects.
Item Operation Description
Projects get_project(project_name) Obtains the name of a MaxCompute project.
exist_project(project_name) Checks whether a MaxCompute project exists.
Tables list_tables() Lists all tables in a MaxCompute project.
exist_table(table_name) Checks whether a table exists.
get_table(table_name, project=project_name) Obtains a specified table. You can obtain a table from another MaxCompute project.
create_table() Creates a table.
read_table() Reads data from a table.
write_table() Writes data to a table.
delete_table() Deletes an existing table.
Table partitions exist_partition() Checks whether a partition exists.
get_partition() Obtains information about a partition.
create_partition() Creates a partition.
delete_partition() Deletes an existing partition.
SQL execute_sql()/run_sql() Executes SQL statements.
open_reader() Reads execution results of SQL statements.
Instances list_instances() Lists all instances in a MaxCompute project.
exist_instance() Checks whether an instance exists.
get_instance() Obtains information about an instance.
stop_instance() Terminates an instance.
Resources create_resource() Creates a resource.
open_resource() Opens a resource.
get_resource() Obtains information about a resource.
list_resources() Lists all existing resources.
exist_resource() Checks whether a resource exists.
delete_resource() Deletes an existing resource.
Functions create_function() Creates a function.
delete_function() Deletes an existing function.
Upload and download tunnels create_upload_session() Creates a session that is used to upload data.
create_download_session() Creates a session that is used to download data.