Apache Kafka Quick Start Guide

Stream processor analysis

If you have a lot of questions here, it is normal.

The first thought to consider is that in streaming aggregation, and in streaming in general, the Streams are unbounded. It is never clear when we will take the final results, that is, we as programmers have to decide when to consider a partial value of an aggregation as a final result.

Recall that the print of the Stream is an instant photo of the KTable at a certain time. Therefore, the results of a KTable are only valid at the time of the output. It is important to remember that in the future, the values of the KTable may be different. Now, to see results more frequently, change the value of the commit interval to zero, shown as follows:

props.put("commit.interval.ms", ...