Grokking the Data Engineering Interview
0% completed
Exactly Once From Stream to Table
Two streaming queries read the same 11 vending machine sales and write a revenue table. Both crash at the same moment, right after writing their second micro-batch, and both are restarted from the same checkpoint. Afterward, one table holds 15 rows and 3,675 cents of revenue, and the other holds 11 rows and the correct 2,700 cents. The engine, the input, and the crash were identical, so the difference comes entirely from how each query writes to its table.
This lesson explains that difference: what a restart replays, and what "exactly once" promises and what it does not
.....
.....
.....
Like the course? Get enrolled and start learning!
Reading Progress
0%