- Introduction of the "Clear Pending Actions" option for Index Unit
- Enhancing Data Query Clarity with Deprecated Property Annotations
User Experience Value
- Unified Indexing Strategy for Index Tabular Pipeline
Data engineers now can mark specific properties within an index unit as "deprecated" by setting an annotation. This annotation serves as a clear signal that certain properties should be used with caution, as they may be phased out or replaced in future updates. This enhances communication between data engineers and users by clearly indicating which properties are deprecated and may be subject to future changes.
With the new release, users now have the "Clear Pending Actions" option in the Index Unit General tab. This option allows users to remove all pending actions targeting the selected index unit. This option provides a quick and efficient way to manage pending actions and streamline subsequent data processing.
With this release, we bring a unified approach to the indexing process across all three formats. Users now have the flexibility to choose their preferred indexing strategy for CSV, JSON, and Parquet files. For all three formats, two choices are available: "mono class" and "multi-class." - Mono Class: users opting for the "mono class" mode can select a single class as the destination for indexing. This choice grants the capability of expression processing before the actual indexing takes place. - Multi-Class: alternatively, users can choose the "multi-class" mode. In this setting, users need only specify the names of the classes in their input file under a designated "class" column. This mode simplifies the indexing process, making it more intuitive and adaptable to diverse data structures.