Uros Stankovic created SPARK-60245:
--------------------------------------
Summary: Support vector_sum and vector_avg to work on double array
for higher precision
Key: SPARK-60245
URL: https://issues.apache.org/jira/browse/SPARK-60245
Project: Spark
Issue Type: New Feature
Components: SQL
Affects Versions: 5.0.0
Reporter: Uros Stankovic
Some mathematical workloads requires higher precision for vector_sum and
vector_avg functions. Those are mathematical functions where return array is
aggregate of all rows of table, and higher precision can benefit workloads such
as Value at risk algorithms.
Other vector functions are used mostly for embedding vectors, and standard
there is array<float>, so this extension is needed only for vector_sum and
vector_avg.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
