[
https://issues.apache.org/jira/browse/CALCITE-6052?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17776428#comment-17776428
]
Mihai Budiu commented on CALCITE-6052:
--------------------------------------
This is a funny one: there is this code in AbstractSqlTester:
{code:java}
@Override public void forEachQuery(SqlTestFactory factory,
String expression, Consumer<String> consumer) {
// Why not return a list? If there is a syntax error in the expression, the
// consumer will discover it before we try to parse it to do substitutions
// on the parse tree.
consumer.accept("values (" + expression + ")");
consumer.accept(buildQuery2(factory, expression));
}
{code}
It turns out that these two never agree for this one test:
{code:java}
f.checkScalarApprox("CAST('-0E0' AS REAL)",
"REAL NOT NULL", "-0E0");
{code}
one of them expects a result of -0.0, while the other expects a result of 0.0.
This signals that there is at least one more bug somewhere
A cowardly solution would be to use checkScalarApprox allowing a rounding
error, but I suspect that's not what we are looking for.
> SqlImplementor writes REAL, FLOAT, or DOUBLE literals as DECIMAL literals
> -------------------------------------------------------------------------
>
> Key: CALCITE-6052
> URL: https://issues.apache.org/jira/browse/CALCITE-6052
> Project: Calcite
> Issue Type: Bug
> Components: core
> Affects Versions: 1.35.0
> Reporter: Mihai Budiu
> Assignee: Mihai Budiu
> Priority: Minor
> Labels: pull-request-available
> Fix For: 1.36.0
>
>
> This bug is already fixed in https://github.com/apache/calcite/pull/3411, but
> I plan to submit a smaller point fix for it, which doesn't require reworking
> the type families.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
