[
https://issues.apache.org/jira/browse/IGNITE-19997?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Iurii Gerzhedovich updated IGNITE-19997:
----------------------------------------
Description:
Currently, we have neither an understanding of how type coercion impacts the
system, nor decent test coverage. As a result, trying to fix one part causes
the other to break (see IGNITE-19128 --> IGNITE-19615 --> IGNITE-19976).
To improve the situation, let us perform the following steps:
* first, it is necessary to cover the current behavior of planner with an
exhaustive set of tests:. Here, we need to cover following cases
** coercion for set operations like UNION, INTERSECT and EXCEPT
** coercion for binary comparison and arithmetic
** coercion for IN operator
** coercion for CASE operator
** coercion for function arguments
** coercion of source for INSERT, UPDATE and MERGE operators
** etc (did I miss something?)
* then, we need to cover execution by the similar tests to make sure coercion
and execution are aligned
* finally, we should revise coercion rules to remove unnecessary casts
was:
Currently, we have neither an understanding of how type coercion impacts the
system, nor decent test coverage. As a result, trying to fix one part causes
the other to break (see IGNITE-19128 --> IGNITE-19615 --> IGNITE-19976).
To improve the situation, let us perform the following steps:
* first, it is necessary to cover the current behavior of planner with an
exhaustive set of tests:. Here, we need to cover following cases
** -coercion for set operations like UNION, INTERSECT and EXCEPT-
** -coercion for binary comparison and arithmetic-
** coercion for IN operator
** coercion for CASE operator
** coercion for function arguments
** coercion of source for INSERT, UPDATE and MERGE operators
** etc (did I miss something?)
* then, we need to cover execution by the similar tests to make sure coercion
and execution are aligned
* finally, we should revise coercion rules to remove unnecessary casts
> Sql. Enhancing test coverage of type coercion
> ---------------------------------------------
>
> Key: IGNITE-19997
> URL: https://issues.apache.org/jira/browse/IGNITE-19997
> Project: Ignite
> Issue Type: Improvement
> Components: sql
> Reporter: Konstantin Orlov
> Priority: Major
> Labels: ignite-3
>
> Currently, we have neither an understanding of how type coercion impacts the
> system, nor decent test coverage. As a result, trying to fix one part causes
> the other to break (see IGNITE-19128 --> IGNITE-19615 --> IGNITE-19976).
> To improve the situation, let us perform the following steps:
> * first, it is necessary to cover the current behavior of planner with an
> exhaustive set of tests:. Here, we need to cover following cases
> ** coercion for set operations like UNION, INTERSECT and EXCEPT
> ** coercion for binary comparison and arithmetic
> ** coercion for IN operator
> ** coercion for CASE operator
> ** coercion for function arguments
> ** coercion of source for INSERT, UPDATE and MERGE operators
> ** etc (did I miss something?)
> * then, we need to cover execution by the similar tests to make sure
> coercion and execution are aligned
> * finally, we should revise coercion rules to remove unnecessary casts
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
