[
https://issues.apache.org/jira/browse/CALCITE-6259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17816489#comment-17816489
]
EveyWu commented on CALCITE-6259:
----------------------------------
hi [~caicancai]
I think we can optimize the log function calculation.
According to the
[document|[https://cloud.google.com/bigquery/docs/reference/standard-sql/mathematical_functions#log]],
the current Caclite calculation for critical conditions does not conform to
the behavior of bigquery.
Such as: LOG(X [, Y]), LOG(0, Y) for bigquery is Error, but for Calcite is
-Infinity； LOG10(0) for bigquery is Error, but for Calcite is -Infinity.
I also think that the background of this issue should be confirmed with
[~tanclary]
> Return the result with the log10 and log functions of argument 0
> ----------------------------------------------------------------
>
> Key: CALCITE-6259
> URL: https://issues.apache.org/jira/browse/CALCITE-6259
> Project: Calcite
> Issue Type: Improvement
> Components: core
> Affects Versions: 1.36.0
> Reporter: Caican Cai
> Priority: Major
> Fix For: 1.37.0
>
> Attachments: 302662660-27b21670-5364-463c-b6dc-d750c46d7cd1.png,
> 302663876-91173a60-695d-409e-b325-3f91655c6d0d.png
>
>
> When log10(0) returns null in mysql and spark, but log10(0) returns error in
> postgres, neither is calcite's -Intity
> {code:java}
> postgres=# select log10(0);
> ERROR: cannot take logarithm of zero
> postgres=# select log(2,0);
> ERROR: cannot take logarithm of zero
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
