Cheng Lian created SPARK-6674:
---------------------------------
Summary: Use different types to represent rows used inside and
outside Catalyst
Key: SPARK-6674
URL: https://issues.apache.org/jira/browse/SPARK-6674
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 1.3.0, 1.2.1, 1.1.1
Reporter: Cheng Lian
Within Catalyst, we use more efficient classes to represent objects in rows.
For example, we use {{o.a.s.sql.types.Decimal}} instead of Java {{BigDecimal}}.
When row objects are passed to / from user space, these object get converted by
{{ScalaReflection.convertToScala}} / {{ScalaReflection.converToCatalyst}}.
This approach is error prone because we may occasionally forget to add
necessary conversions. It would be better to have a dedicated row type to
represent row objects used within Catalyst, thus enforce this convention via
type system.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
