Yes, I think you are in the right direction. Vignesh <[email protected]> 于2023年2月8日周三 05:31写道：
> Hi, > > To do this, I will need to define an implementation of > CalciteCatalogReader, override the getTable method and return an > implementation of PreparingTable that returns the right thing for > getMonotonicity > < > https://javadoc.io/static/org.apache.calcite/calcite-core/1.11.0/org/apache/calcite/sql/validate/SqlValidatorTable.html#getMonotonicity(java.lang.String) > >and > supportsModality > < > https://javadoc.io/static/org.apache.calcite/calcite-core/1.11.0/org/apache/calcite/sql/validate/SqlValidatorTable.html#supportsModality(org.apache.calcite.sql.validate.SqlModality) > > > . > Is above the right approach? Would there be more steps beyond above? > > Thanks, > Vignesh. > > On Mon, Jan 30, 2023 at 10:50 PM Vignesh <[email protected]> wrote: > > > Thank you, I will try that. > > > > On Sun, Jan 29, 2023, 9:01 PM Benchao Li <[email protected]> wrote: > > > >> Hi Vignesh, > >> > >> AFAIK, there is no such way in Calcite to declare it with DDL. However, > >> you > >> can provide it via the statistic for your table, you can take a look at > >> the > >> tests[1] in Calcite how we did it. > >> > >> [1] > >> > >> > https://github.com/apache/calcite/blob/ae228f64347cc7620d28eff04f5869583c6bf9d5/testkit/src/main/java/org/apache/calcite/test/catalog/MockCatalogReaderSimple.java#L238-L245 > >> > >> Vignesh <[email protected]> 于2023年1月30日周一 06:21写道： > >> > >> > Hi, > >> > > >> > This page <https://calcite.apache.org/docs/stream.html> describes > >> > streaming > >> > queries in calcite. > >> > Streaming operators, such as tumbling window require a monotonic > >> column. It > >> > says that we need to declare monotonic columns in the schema. > >> > > >> > Is there a way to do it in the create table statement? Flink SQL does > it > >> > with a "WATERMARK FOR rowtime_column" > >> > < > >> > > >> > https://nightlies.apache.org/flink/flink-docs-master/docs/dev/table/sql/create/#create-table > >> > > > >> > syntax in the create table, but that seems to be flink specific. > >> > > >> > > >> > Thanks, > >> > Vignesh. > >> > > >> > >> > >> -- > >> > >> Best, > >> Benchao Li > >> > > > -- Best, Benchao Li
