[
https://issues.apache.org/jira/browse/CALCITE-3760?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17026943#comment-17026943
]
Julian Hyde edited comment on CALCITE-3760 at 1/30/20 7:22 PM:
---------------------------------------------------------------
SQL is a [strict
language|https://en.wikipedia.org/wiki/Strict_programming_language] (with the
exception of a few constructs such as CASE) but becomes non-strict when you add
non-deterministic UDFs. As you point out, some of our rewrites assume
strictness. It would be helpful if we had a 'let' construct, e.g.
{{coalesce(e1, e2)}} becomes {{let v = e1 in case when x is not null then x
else e2 end}}. It would ensure that expressions are evaluated the correct
number of times. Without {{let}} or something similar I don't know how we could
do these rewrites.
Related issues are re-ordering of the branches of AND and OR conditions, and
behavior when an expression throws.
was (Author: julianhyde):
SQL is a [strict
language|https://en.wikipedia.org/wiki/Strict_programming_language] (with the
exception of a few constructs such as CASE) but becomes non-strict when you add
non-deterministic UDFs. As you point out, some of our rewrites assume
strictness. It would be helpful if we had a 'let' construct, e.g.
{{coalesce(e1, e2)}} becomes {{let v = e1 in case when x is not null then x
else e2 end}}. It would ensure that expressions are evaluated the correct
number of times. Without {{let}} or something similar I don't know how we could
do these rewrites.
> Rewriting non-deterministic function can break query semantics
> --------------------------------------------------------------
>
> Key: CALCITE-3760
> URL: https://issues.apache.org/jira/browse/CALCITE-3760
> Project: Calcite
> Issue Type: Bug
> Components: core
> Reporter: Jin Xing
> Assignee: Jin Xing
> Priority: Major
>
> Calcite rewrite some *SqlFunctions* during validation. But whether the
> function is deterministic is not considered. For a non-deterministic
> operator, the rewriting can break semantics. Additionally there's no
> interface for user to specify the determinism for a UDF/UDAF.
> Say I have non-deterministic UDF & UDAF and run sql like below
> {code:java}
> select coalesce(udf(col0), 100) from foo;
> select nullif(udaf(col0), 1024) from foo;{code}
> They will be rewritten as
> {code:java}
> select case when udf(col0) is not null then udf(col0) else 100 end
> from foo;
> select case when udaf(col0)=1024 then null udaf(col0)
> from foo{code}
> As we can see that non-deterministic UDF & UDAF are called multiple times
> after written. Thus the condition in WHEN clause might NOT be held all the
> time.
> We need to provide an interface for user to specify the determinism in
> UDF/UDAF and consider whether a SqlNode is deterministic when rewriting.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
