Unstructured data processing functions

Updated at:

MaxCompute provides multiple ways to access unstructured data and its metadata from a data warehouse or data lake.

Object Table and OSS object functions

MaxCompute supports the Object Table feature to access unstructured data and its metadata in OSS. It also provides functions to process this data. For more information, see Object Table definition.

Function

Description

GET_DATA_FROM_OSS

Reads all or part of an object and returns the content in binary format.

GET_SIGNED_URL_FROM_OSS

Generates a pre-signed URL to download data from or upload data to OSS. This allows you to read and write OSS objects directly over HTTP.

GET_BLOB_FROM_OSS

Generates a Blob value that points to all or part of an OSS object, based on the object key recorded in an Object Table.

Blob object processing functions

MaxCompute supports the Blob (binary large object) data type. This data type is designed for storing unstructured or semi-structured binary data, such as images, file content, and serialized data. The following functions are available for processing Blob objects.

Function

Description

GET_BLOB_METADATA

Reads the metadata of a Blob. You can then use a JSON function to extract specific fields from the returned value.