Google Cloud Platform Certification - Data Engineer - Exam Review- Prepping for exam

sanguru (35)in #google • 7 years ago

Personally I thought the exam was written in a very unprocessed approach and appeared rushed
Did not appear to use best practices in exam development such as Bloom.
Case studies. I did like how they have the case studies listed on Exam Guide so you could review before taking exam and not spend time during exam reviewing.
Exam did have a technical merit but as a routine test developer I see the need for a better exam guide and test JTA to be completed.

Case studies were part of the exam and you needed to review and answer the appropriate solution for the specific questions. Case study had numerous questions similar but had a slight question or answers so you needed to pay attention.
Cloud DataProc - Questions about migrating onsite Hadoop and how Cloud DataProc could help.
Cloud DataFlow - Numerous questions around knowing how Cloud DataFlow fits, know stream vs batch, it’s a managed service so that you don’t need to deal with it, how you can use for ETL. Lastly, you will see several questions related to how Cloud Dataflow can manages services in Google Cloud Storage and Google Cloud Compute Engine. (Yes, know this)

Pipelines in DataFlow and how you could use graph objects. One question about why you would use JSON or Java related to pipelines.
Storage around every aspect and needed to discern between Nearline and Coldline. Big Data , Regional and Standard storage. Cloud Storage is a must to know.
Hadoop. Numerous questions around HDFS but also when you would need to use approaches like Hive, Sqoop, Oozie with Hadoop.
Stackdriver –not really what I would call Data Engineering but they asked four questions about this Hybrid monitoring service . I remember a few questions focused on how you can debug, monitor and log using Stackdriver. I think they were checking to confirm if you knew how use Stackdriver to help debug source code, etc.

BigTable- numerous questions about why and how you could use BigTable. Know it’s a high performance NoSQL database service for large analytical and operational workloads.
BigQuery. Tested heavily around data ingestion and availability. Once again they want to confirm you know that it’s a fully managed, petabyte scale, low cost enterprise data warehouse for analytics. BigQuery is serverless.
Cloud SQL. They wanted to confirm you knew it and why over other services like BigTable, BigQuery, etc.