Dataflow & Apache Beam interview notes for GCP roles
1. What is a PCollection?
An immutable, distributed data set in a Beam pipeline – bounded (batch) or unbounded (streaming).
2. ParDo vs Map?
ParDo is the general element-wise transform that can emit zero or more outputs per input; Map is the simple one-in, one-out form.
3. What is a side input?
A small, extra data set (for example a lookup table) made available to every worker while processing the main input.
Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle. Additional detail on autoscaling, fusion and shuffle.
Feedback & comments (0)
Be the first to share feedback on this post.