Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Developer practice bank has an additional 608 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team is developing a microservice that collects continuous telemetry events from millions of connected devices and ingests them into BigQuery. Your ingestion architecture must meet the following requirements:
Which ingestion pattern should you implement?
The default stream in the BigQuery Storage Write API (gRPC) is a built-in, pre-created stream available for every BigQuery table. It enables applications to stream records directly to a table without explicitly creating or managing individual write stream resources, offering at-least-once delivery semantics optimized for maximum write throughput.
Your workload prioritized maximum ingestion throughput and simplicity over strict deduplication, and downstream analytical queries could tolerate occasional duplicate events or handle deduplication during post-processing.
Developers often assume that exponential backoff in client libraries prevents duplicates. In reality, retries only protect against lost writes; client-provided offsets on an explicit committed stream are required to prevent duplicate appends during retry cycles.
BigQuery batch loading (invoked via bq load, the LOAD DATA SQL statement, or client libraries) ingests static files—such as newline-delimited JSON (ndJSON), CSV, Avro, Parquet, or ORC—from Cloud Storage into a BigQuery table as a single atomic job. During ingestion, BigQuery converts the source files into its Capacitor columnar storage format.
gzip cannot be read in parallel and are limited to 4 GB per file.Your workload accumulated telemetry into large, scheduled batches (such as hourly or daily files in Cloud Storage) where zero ingestion cost is prioritized over real-time query availability.
When data arrives continuously with frequent updates, BigQuery best practices recommend streaming via the Storage Write API (gRPC) specifically to avoid hitting load job quota limits while maintaining low latency.
Event-driven transfers in the BigQuery Data Transfer Service automate incremental batch loading from Cloud Storage into BigQuery. By configuring Cloud Storage bucket notifications for OBJECT_FINALIZE events on a Pub/Sub topic and linking a transfer configuration to a dedicated subscription, BigQuery automatically schedules load jobs whenever new or modified files appear in the bucket.
OBJECT_FINALIZE messages arrive immediately.updated timestamps to avoid reloading unchanged files in APPEND mode, though modifying files mid-transfer can still cause inconsistencies.Your application produced periodic file drops in Cloud Storage and could tolerate a several-minute delay (up to 10 minutes) before data became queryable in BigQuery, prioritizing low-cost managed batch automation over real-time streaming.
Staging files in Cloud Storage is ideal for ELT batch pipelines, but whenever real-time visibility and minimal latency are required, streaming directly through the Storage Write API (gRPC) is the recommended pattern.
The BigQuery Storage Write API (gRPC) is a unified, high-throughput streaming ingestion protocol designed for real-time data pipelines. When configured with a committed write stream, records sent via AppendRows calls become immediately visible for querying as soon as they are acknowledged by BigQuery, while supporting client-provided stream offsets to guarantee atomic, ordered writes.
google-cloud-bigquery-storage) automatically apply exponential backoff and jitter for transient gRPC UNAVAILABLE errors.UPDATE, DELETE, and MERGE statements on data streamed within the last 30 minutes.BigQueryIO connector for scalable streaming pipelines.Unlike batch transfers from Cloud Storage or offset-free streaming, a committed write stream with client offsets is the only BigQuery ingestion mechanism that simultaneously delivers sub-second query visibility, strict exactly-once guarantees across retries, and independence from batch load job quotas.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated