Hi, all !
Whats the purpose for RelJson(Reader|Writer) classes, is it just an
example how RelNode`s can be serialized?
If so - it answers my question.
But if it something more seems there is some bugs are present after [1]
was implemented,
In particular NaN and Inf cannot be correctly deserialized, short
reproducer below, do i need to fill an issue?
Thanks !
place this into RelWriterTest
---
@Test void testInf() {
Frameworks.withPlanner((cluster, relOptSchema, rootSchema) -> {
final RexBuilder rexBuilder = cluster.getRexBuilder();
final FrameworkConfig config = RelBuilderTest.config().build();
final RelBuilder builder = RelBuilder.create(config);
RelDataType doubleType =
cluster.getTypeFactory().builder().add("name",
cluster.getTypeFactory().createSqlType(SqlTypeName.DOUBLE)).build();
final RelDataType doubleTypeLit =
cluster.getTypeFactory().createSqlType(SqlTypeName.DOUBLE);
RexLiteral lit =
rexBuilder.makeApproxLiteral(Double.POSITIVE_INFINITY, // <---
doubleTypeLit);
RelNode vals = builder.values(Collections.nCopies(1,
List.of(lit)), doubleType).build();
final RelJsonWriter writer = new RelJsonWriter();
vals.explain(writer);
String serialized = writer.asString();
SchemaPlus schema = rootSchema.add("hr",
new ReflectiveSchema(new HrSchema()));
final RelJsonReader reader =
new RelJsonReader(cluster, relOptSchema, schema);
RelNode node;
try {
node = reader.read(serialized); // <- raised from here
} catch (IOException e) {
throw TestUtil.rethrow(e);
}
return RelOptUtil.dumpPlan("", node, SqlExplainFormat.TEXT,
SqlExplainLevel.EXPPLAN_ATTRIBUTES);
});
}
---
will raise:
com.fasterxml.jackson.core.JsonParseException: Non-standard token
'Infinity': enable `JsonReadFeature.ALLOW_NON_NUMERIC_NUMBERS` to allow
[1] https://issues.apache.org/jira/browse/CALCITE-2067
