Performance Optimization Flashcards

Which microbatch behavior makes the incremental strategy more efficient for large time-series data?

Answer: Processing data in separate time-based batches that can be retried or run in parallel

Microbatch splits the build into batches by event_time so each batch can be processed, retried and backfilled independently.

In dbt Cloud, what does a state-aware approach to jobs primarily avoid?

Answer: Rebuilding models whose upstream data and code have not changed

State-aware orchestration skips nodes with no changes in code or upstream data, saving warehouse compute.

Which command checks whether source data is stale, so downstream builds can be skipped when nothing new arrived?

Answer: dbt source freshness

dbt source freshness compares loaded_at_field against thresholds and writes sources.json.

Which practice best keeps an incremental model from scanning the entire source on every run?

Answer: Wrapping a filter on a timestamp column inside an is_incremental() block

Filtering new rows against the max timestamp in the target inside is_incremental() limits the data read each run.

What is the performance benefit of selecting only required columns instead of SELECT * in a columnar warehouse?

Answer: Reduced bytes scanned and lower cost

Columnar stores read only referenced columns, so explicit selection lowers I/O and cost.

What does the dbt build command do differently from running run and test separately?

Answer: Builds seeds, snapshots, models and tests in DAG order, stopping downstream nodes on test failure

dbt build executes resources in dependency order and skips downstream nodes when an upstream test fails.

Which Redshift config influences how table rows are distributed across nodes to reduce data shuffling in joins?

Answer: dist

The dist config sets the distribution style or key for Redshift tables, minimizing data movement during joins.