Details
-
New Feature
-
Status: Open
-
Not a Priority
-
Resolution: Unresolved
-
1.14.4
-
None
Description
When we run Flink SQL using upsert-kafka connector, it cannot support startup mode of timestamp or specific offsets. It is determined to read data from the earliest.
I don't think it's reasonable. Because the data stored in Kafka is incomplete that it only saves data for 7 days in most cases. So why don't we consume data from the specified timestamp or offsets, and this can improve the efficiency of reading data from Kafka.
Attachments
Issue Links
- links to