2024)¶

Added read support for Glue Catalogs in BodoSQL
Added support for more DDL commands for both Snowflake and Iceberg:
- DESCRIBE VIEW (requires view support in catalog)
- DROP VIEW (requires view support in catalog)
- ALTER TABLE ... RENAME TO
- ALTER VIEW ... RENAME TO (requires view support in catalog)
- SHOW TERSE TABLES
- SHOW TERSE VIEWS (requires view support in catalog)
Added support for pushing down join filters into cache nodes and uses join filters to determine whether or not to keep cache nodes.
Added support for casting strings to Decimal types

Allowed limited support for streaming execution of window functions, particularly for DENSE_RANK.
Allowed Join Filters to be pushed into Snowflake I/O calls, particularly when the build side of a join allows us to infer the minimum/maximum value of a join key.

Allowed Tabular Catalog to use testing environment
Handled trailing slashes in Tabular Catalog's REST URI when created through the Bodo Platform
Enabled further decorrelation with builtin table functions

2024.6.1¶

Improves the memory manager to reduce the likelihood of out-of-memory errors in certain situations.
Improved NDV estimates involving CASE expressions and IS NOT NULL filters.
Runtime join filters are now applied in the order that's most likely to reduce intermediate costs.

Fixed ALTER TABLE RENAME TO not being able to specify schemas or databases.
TIMESTAMPTZ can be compared with other TIMESTAMP types in all contexts
Fixed calling GROUP BY with keys that mixed LIST columns with other types
Automatically casts the decimal input to a floating point type when computing DECIMAL * FLOAT