Calcite is counter intuitive. It should be in
reverse order, first generate a semi or anti-semi join, then generate an
inner/outer join.
- Haisheng
--
发件人：Christian Beikov
日 期：2020年02月19日 21:12:13
收件人：
主 题：Translation of SQL EXISTS
semi join, then generate an
>>> inner/outer join.
>>>
>>> - Haisheng
>>>
>>> --
>>> 发件人：Christian Beikov
>>> 日 期：2020年02月19日 21:12:13
>>> 收件人：
>>>
this process in Calcite is counter intuitive. It should be in
reverse order, first generate a semi or anti-semi join, then generate an
inner/outer join.
- Haisheng
--
发件人：Christian Beikov
日 期：2020年02月19日 21:12:13
收件人：
主 题：Translation
---
> 发件人：Christian Beikov
> 日 期：2020年02月19日 21:12:13
> 收件人：
> 主 题：Translation of SQL EXISTS
>
> Hello,
>
> I'm a bit confused about how the SQL EXISTS predicate is translated. I'd
> assume that an EXISTS is translated in relational algebra to a S
--
发件人：Christian Beikov
日 期：2020年02月20日 03:11:02
收件人：Haisheng Yuan; Apache Calcite dev
list
主 题：Re: Translation of SQL EXISTS
Hey Haisheng,
it is nice to have a rule that detects such patterns but do you agree that it
would be better to generate
：Translation of SQL EXISTS
Hello,
I'm a bit confused about how the SQL EXISTS predicate is translated. I'd
assume that an EXISTS is translated in relational algebra to a SEMI- and
NOT EXISTS to an ANTI-join, but it's not.
PlannerImpl p = new PlannerImpl(config);
SqlNode
：Christian Beikov
日 期：2020年02月19日 21:12:13
收件人：
主 题：Translation of SQL EXISTS
Hello,
I'm a bit confused about how the SQL EXISTS predicate is translated. I'd
assume that an EXISTS is translated in relational algebra to a SEMI- and
NOT EXISTS to an ANTI-join, but it's not.
PlannerImpl p = new
Hello,
I'm a bit confused about how the SQL EXISTS predicate is translated. I'd
assume that an EXISTS is translated in relational algebra to a SEMI- and
NOT EXISTS to an ANTI-join, but it's not.
PlannerImpl p = new PlannerImpl(config);
SqlNode sqlNode = p.parse("delete from _order o where
