Tables
Table-based Offline Stores, such as Snowflake, inherently revolve around tables. These providers furnish a method known as.register_table(name, variant, table="")
. This method enables you to register transformations based on the primary data set or to register features and labels derived from it.
Files
Offline Stores like Spark interact with file stores like S3 and HDFS. For these providers, there exists a method called.register_file(name, variant="", path="")
.
Currently, Featureform offers support for CSVs and Parquet files. If your specific use case requires a different file format, please don’t hesitate to raise an issue on Github or engage with our community on Slack. We value your feedback and are eager to explore new possibilities.