Bodo 2023.6 Release (Date: 06/23/2023)¶
New Features and Improvements¶
Bodo:
- Added original date type support for all datetime functions.
- Support filter pushdown for various functions.
- Upgrade to Arrow 11.
- Improved performance on join operations by using streaming loop.
- Support nullable timestamp, float and boolean array.
- Zero-Copy support for most Arrow Arrays.
- Supported passing timezone to pd.Timestamp.now().
- Support for comparison operators between date and tz-aware/naive timestamps.
BodoSQL:
Added support for the following functions:
- CURRENT_DATE
- DATEDIFF/TIMEDIFF/TIMESTAMPDIFF
- TRY_CAST
- LEAD/LAG
- [TRY_]TO_BINARY
- [TRY_]TO_DECIMAL, [TRY_]TO_NUMBER, [TRY_]TO_NUMERIC, [TRY_]TO_DOUBLE
- [TRY_]TO_DOUBLE
- [TRY_]TO_TIME
- SAMPLE
- ILIKE/ANY/ALL
- LEAST/GREATEST
- ADD_MONTH/MONTH_BETWEEN
- HASH
- RANDOM
- UNIFORM
- TO_ARRAY
- ARRAY_TO_STRING
- SPLIT
Parity Improvements:
- Added support for reading nested arrays.
- Support
repeats
for tz-aware data in join optimization. - Support logical and comparison operators between Boolean and Numeric Types.
- Support for date outputs in Case statements.
Other Improvements:
- More efficient fill templating for Join.
- Improved Snowflake sampling for dict-encoding detection.
- Revamped C++ array representation to be more robust.
- Support batching / streaming Snowflake Read Implementation.