Data assets and MLTable in Azure Machine Learning

A data asset is a named, versioned reference to data, built on a URI that points at a datastore or storage endpoint. The URI scheme must match the storage endpoint, and the asset type must match what the path points at.

Common trap: Using abfss for a data asset over all the blobs in a container - abfss needs the ADLS Gen2 dfs endpoint; for the blob endpoint, use wasbs with URI_FOLDER.

• An MLTable file is a YAML blueprint that describes how to read the data files in its folder. Pass the containing folder to mltable.load("./iris"), then call to_pandas_dataframe() to get the data. save() writes a blueprint, take() returns the first rows, and take_random_sample() returns a random sample.

Common trap: Passing the MLTable file itself to load(), as in mltable.load("./sample_data/MLTable") - load() takes the folder that contains the MLTable file, not the file.

Get the whole book

This note is one section of Ultra Transcenders AI-300: Operationalizing Machine Learning and Generative AI Solutions, an independent study guide that explains every topic the exam covers by technology, with comparison tables, diagrams and the common traps, plus a glossary linked to Microsoft Learn.

About the book · Free AI-300 glossary · All AI-300 study notes