
Updated OpenSSL dependency from 1.1.1g to 1.1.1k. This problem did not affect the default file format. Fixed a problem that could corrupt databases in rare cases when the new file format has explicitly been enabled by the user. Java: Dependencies are updated to newer versions. However, the Hyper API will continue to support Python 3.6 for a transition period of three months. Currently, those versions are 3.7, 3.8, and 3.9. Going forward, the Hyper API will only support the three most recent versions of Python. More actionable error messages for a wide range of invalid SQL queries. Column references can now be qualified with database alias (e.g., SELECT db. Parquet files with dots in their column names can now be read. Upgraded Unicode support from Unicode 9.0.0 to 13.0.0. Fix a query compilation defect that led to reproducible crashes for a very small number of queries. Updated OpenSSL version from 1.1.1k to 1.1.1l. Fixed a defect that could cause crashes when you deleted a tuple multiple times, or could cause wrong query results after single deletion. calls without PARTITION BY and ORDER BY clauses, e.g.
calls without PARTITION BY clause improved by 5% - 20%.Improved performance for window functions for large data sets on multi-core machines:.Reading Apache Parquet files is now officially supported and no longer has to be enabled through the use of the process setting ‘experimental_external_format_parquet’.* Rename Pacific/Enderbury to Pacific/Kanton. * Merge more location-based Zones whose timestamps agree since 1970. * Jordan now starts DST on February's last Thursday. * Fiji suspends DST for the 2021/2022 season. Noteworthy changes in the Time Zone Database:
This commit updates Hyper’s Time Zone Database to release 2021e.
IANA released new versions of the Time Zone Database. Fixed a defect that could lead to Hyper sorting and comparing text with the “ro” locale incorrectly. Fixed a defect that could lead to crashes when reading Parquet files with text columns that contain null values.
Java: Improved the read performance of text and geography columns.
#Install tableau hyper api java download
The download size of a packed Windows Python package was reduced from 47 MB to 42 MB. For example, the size of an unpacked Windows Python installation went from 157 MB to 145 MB.
Smaller packages: Thanks to the removal of unused collation data from the Hyper binary, the package size was reduced.