Pre-Release of Polars 2.0

Polars has released a pre-release candidate of version 2.0, with the official 2.0 release expected in the coming weeks. The primary focus of this major version upgrade is to streamline the software and improve user experience. The most significant change is the default shift from the in-memory engine to the streaming engine for all LazyFrame queries, promising significant improvements in memory usage and performance.

Users can expect up to 5 times faster processing with the streaming engine. To assist with the transition, a comprehensive migration guide has been provided. The upgrade also introduces stricter error handling, such as raising exceptions for data-type mismatches and removing ambiguous casts. Users can opt-in to maintain row order in certain operations, like join or group_by, by setting maintain_order=True.

Polars aims to fail fast, providing clear error messages to help users quickly identify and correct issues. While most removed functionality had been deprecated previously, the new exceptions will guide users to the updated API.