Improved column ordering in schema inference for TSKV and JSONEachRow formats, with enhanced handling of empty rows during schema detection.
Why it matters
This feature addresses issues in schema inference for the TSKV and JSONEachRow input formats by ensuring that columns are ordered consistently and that the inference process does not stop when encountering empty rows. This improves data loading reliability and correctness when processing these formats.How to use it
Users benefit from this improvement automatically when using the TSKV or JSONEachRow formats with schema inference enabled. No special configuration is required; the system now handles empty rows gracefully and infers schema columns in a better order.