Hi Eugenio,
According to docs[1], there are two ways to define the watermark in a table:
1. Defining in DDL
2. During DataStream-to-Table Conversion
In your case, I think could use CREATE TABLE DDL to create a new table from
filteredPhasesDurationsTable with watermark.
See more in CREATE Statemen
Hi, Eugenio
AFAIK, you could define watermark on the data_fine by adding attribute
in phasesDurationsSchema.
For example:
final Schema phasesDurationsSchema = Schema.newBuilder()
.column("id_fascicolo", DataTypes.BIGINT().notNull())
.column("nrg", DataTypes.STRING())
.column("giudice", DataTypes.
Hi everyone,
I'm using Flink for processing some streaming data. First of all I have two
tables receiving events from Kafka. These tables are joined and the resulting
table is converted to a DataStream where it is processed by a custom
KeyedProcessFunction. The output is then converted to a tab
Hi,
>From the attachment, I saw there are IntervalJoin and GroupWindowAggregate
operators. AFAIK the state retention for such operators is not controlled
by `table.exec.state.ttl`.
Could you share the operator-level state metrics to help identify the
issue?
Best,
Jane
On Sun, Jun 25, 2023 at 10:
