machine learning engineer - associate - data types

data types supported by ML

what are the column-based data formats for ML and when are they used?

Column-based data formats, such as Parquet and ORC, are optimized for read-heavy analytical workloads and are often used in big data processing frameworks. They store data in a columnar layout, which allows for efficient compression and querying of specific columns.

what are the row-based formats for ML and when are they used?

Row-based formats, such as CSV and avro RecordIO, are typically used for transactional workloads and are optimized for write-heavy operations. They store data in a row-wise layout, making it easier to read and write entire records efficiently.

what are the record type formats used in ML?

Record type formats, such as JSON, JSONL and avro RecordIO, are used for data interchange and serialization. They allow for flexible schema definitions and are suitable for both structured and semi-structured data.

what are the object notation data types for ML and when are they used?

JSON and JSONL are used for non-tabular hierarchical data