dbt Advanced interview questions

1. A.A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.
2. B.A dbt config that enforces a model's declared column names, data types, and constraints at build time, failing if the output does not match.
3. C.A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
4. D.A 2025 ground-up rewrite of the dbt engine in Rust with native SQL comprehension, offering much faster parsing and compilation than dbt Core.

Answer + AI explanation with Pro

1. A.pre-hook and post-hook
2. B.dbt package
3. C.dbt Fusion
4. D.Semantic Layer

Answer + AI explanation with Pro

1. A.dbt Fusion — An incremental-model config controlling how dbt reacts to source column changes, with options ignore, fail, append_new_columns, and sync_all_columns.
2. B.dbt Fusion — A 2025 ground-up rewrite of the dbt engine in Rust with native SQL comprehension, offering much faster parsing and compilation than dbt Core.
3. C.dbt Fusion — A dbt test that validates a model's SQL logic against small fixed mock inputs and expected output, run without touching warehouse data.
4. D.dbt Fusion — A dbt config that enforces a model's declared column names, data types, and constraints at build time, failing if the output does not match.

Answer + AI explanation with Pro

1. A.A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
2. B.An incremental strategy that deletes target rows matching the new batch's keys and then inserts the batch, used on Snowflake and Redshift.
3. C.An incremental strategy that replaces whole partitions of the target with new data rather than row-level merging, common on BigQuery and Spark.
4. D.A pattern for splitting a monolithic dbt project into multiple governed projects that reference each other across boundaries.

Answer + AI explanation with Pro

1. A.unit test
2. B.model contract
3. C.delete+insert strategy
4. D.dbt Mesh

Answer + AI explanation with Pro

1. A.dbt Mesh — A pattern for splitting a monolithic dbt project into multiple governed projects that reference each other across boundaries.
2. B.dbt Mesh — A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
3. C.dbt Mesh — An incremental strategy that deletes target rows matching the new batch's keys and then inserts the batch, used on Snowflake and Redshift.
4. D.dbt Mesh — An incremental strategy that replaces whole partitions of the target with new data rather than row-level merging, common on BigQuery and Spark.

Answer + AI explanation with Pro

1. A.A dbt model that captures slowly changing dimension history by recording row versions over time into a snapshot table.
2. B.A dbt config that enforces a model's declared column names, data types, and constraints at build time, failing if the output does not match.
3. C.A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
4. D.The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.

Answer + AI explanation with Pro

1. A.access modifier
2. B.MetricFlow
3. C.insert_overwrite strategy
4. D.on_schema_change

Answer + AI explanation with Pro

1. A.access modifier — A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.
2. B.access modifier — A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
3. C.access modifier — A dbt node that documents a downstream use of project data (a dashboard or app), making it a tracked dependency in lineage and freshness.
4. D.access modifier — A dbt test that validates a model's SQL logic against small fixed mock inputs and expected output, run without touching warehouse data.

Answer + AI explanation with Pro

1. A.A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.
2. B.A snapshot strategy that detects changes by comparing a chosen set of columns for any difference, used when no reliable updated_at exists.
3. C.The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.
4. D.A snapshot strategy that detects changes by comparing an updated_at column, treating a newer timestamp as a changed row.

Answer + AI explanation with Pro

1. A.dbt package
2. B.check snapshot strategy
3. C.Semantic Layer
4. D.pre-hook and post-hook

Answer + AI explanation with Pro

1. A.Semantic Layer — A dbt config that enforces a model's declared column names, data types, and constraints at build time, failing if the output does not match.
2. B.Semantic Layer — A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.
3. C.Semantic Layer — A pattern for splitting a monolithic dbt project into multiple governed projects that reference each other across boundaries.
4. D.Semantic Layer — A dbt check comparing the max loaded-at timestamp of a source table against warn_after and error_after thresholds to detect stale upstream data.

Answer + AI explanation with Pro

1. A.A snapshot strategy that detects changes by comparing a chosen set of columns for any difference, used when no reliable updated_at exists.
2. B.The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.
3. C.A dbt node that documents a downstream use of project data (a dashboard or app), making it a tracked dependency in lineage and freshness.
4. D.Using a stored manifest with the state: selector and dbt build --select state:modified to run only changed models in CI (slim CI).

Answer + AI explanation with Pro

1. A.MetricFlow
2. B.source freshness
3. C.pre-hook and post-hook
4. D.timestamp snapshot strategy

Answer + AI explanation with Pro

1. A.MetricFlow — A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
2. B.MetricFlow — An incremental strategy that upserts using a unique_key, updating matched rows and inserting new ones in a single MERGE statement.
3. C.MetricFlow — Using a stored manifest with the state: selector and dbt build --select state:modified to run only changed models in CI (slim CI).
4. D.MetricFlow — The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.

Answer + AI explanation with Pro

1. A.The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.
2. B.Using a stored manifest with the state: selector and dbt build --select state:modified to run only changed models in CI (slim CI).
3. C.An incremental strategy that replaces whole partitions of the target with new data rather than row-level merging, common on BigQuery and Spark.
4. D.A dbt test that validates a model's SQL logic against small fixed mock inputs and expected output, run without touching warehouse data.

Answer + AI explanation with Pro

1. A.timestamp snapshot strategy
2. B.snapshot
3. C.dbt state comparison
4. D.check snapshot strategy

Answer + AI explanation with Pro

1. A.dbt state comparison — Using a stored manifest with the state: selector and dbt build --select state:modified to run only changed models in CI (slim CI).
2. B.dbt state comparison — The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.
3. C.dbt state comparison — A 2025 ground-up rewrite of the dbt engine in Rust with native SQL comprehension, offering much faster parsing and compilation than dbt Core.
4. D.dbt state comparison — A dbt model that captures slowly changing dimension history by recording row versions over time into a snapshot table.

Answer + AI explanation with Pro

1. A.An incremental strategy that deletes target rows matching the new batch's keys and then inserts the batch, used on Snowflake and Redshift.
2. B.A dbt check comparing the max loaded-at timestamp of a source table against warn_after and error_after thresholds to detect stale upstream data.
3. C.A snapshot strategy that detects changes by comparing a chosen set of columns for any difference, used when no reliable updated_at exists.
4. D.A dbt model that captures slowly changing dimension history by recording row versions over time into a snapshot table.

Answer + AI explanation with Pro

1. A.timestamp snapshot strategy
2. B.source freshness
3. C.dbt group
4. D.merge strategy

Answer + AI explanation with Pro

1. A.source freshness — A dbt config that enforces a model's declared column names, data types, and constraints at build time, failing if the output does not match.
2. B.source freshness — SQL statements dbt runs immediately before or after building a model, used for grants, vacuuming, or auditing.
3. C.source freshness — Using a stored manifest with the state: selector and dbt build --select state:modified to run only changed models in CI (slim CI).
4. D.source freshness — A dbt check comparing the max loaded-at timestamp of a source table against warn_after and error_after thresholds to detect stale upstream data.

Answer + AI explanation with Pro

1. A.An incremental strategy that upserts using a unique_key, updating matched rows and inserting new ones in a single MERGE statement.
2. B.A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.
3. C.An incremental-model config controlling how dbt reacts to source column changes, with options ignore, fail, append_new_columns, and sync_all_columns.
4. D.A dbt model that captures slowly changing dimension history by recording row versions over time into a snapshot table.

Answer + AI explanation with Pro

1. A.delete+insert strategy
2. B.merge strategy
3. C.dbt group
4. D.snapshot

Answer + AI explanation with Pro

1. A.snapshot — A 2025 ground-up rewrite of the dbt engine in Rust with native SQL comprehension, offering much faster parsing and compilation than dbt Core.
2. B.snapshot — A dbt model that captures slowly changing dimension history by recording row versions over time into a snapshot table.
3. C.snapshot — The dbt engine that compiles semantic metric definitions into correct warehouse SQL, handling joins, aggregations, and time grains.
4. D.snapshot — A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.

Answer + AI explanation with Pro

1. A.A 2025 ground-up rewrite of the dbt engine in Rust with native SQL comprehension, offering much faster parsing and compilation than dbt Core.
2. B.A pattern for splitting a monolithic dbt project into multiple governed projects that reference each other across boundaries.
3. C.A dbt test that validates a model's SQL logic against small fixed mock inputs and expected output, run without touching warehouse data.
4. D.A snapshot strategy that detects changes by comparing an updated_at column, treating a newer timestamp as a changed row.

Answer + AI explanation with Pro

1. A.MetricFlow
2. B.source freshness
3. C.timestamp snapshot strategy
4. D.delete+insert strategy

Answer + AI explanation with Pro

1. A.timestamp snapshot strategy — A dbt model that captures slowly changing dimension history by recording row versions over time into a snapshot table.
2. B.timestamp snapshot strategy — A dbt node that documents a downstream use of project data (a dashboard or app), making it a tracked dependency in lineage and freshness.
3. C.timestamp snapshot strategy — A 2025 ground-up rewrite of the dbt engine in Rust with native SQL comprehension, offering much faster parsing and compilation than dbt Core.
4. D.timestamp snapshot strategy — A snapshot strategy that detects changes by comparing an updated_at column, treating a newer timestamp as a changed row.

Answer + AI explanation with Pro

1. A.A reusable bundle of models, macros, and tests installed via packages.yml, such as dbt_utils, that is imported into a project.
2. B.A snapshot strategy that detects changes by comparing a chosen set of columns for any difference, used when no reliable updated_at exists.
3. C.A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
4. D.An ownership boundary that assigns models to a team and can restrict access, enabling governance across a large mesh of projects.

Answer + AI explanation with Pro

1. A.source freshness
2. B.check snapshot strategy
3. C.pre-hook and post-hook
4. D.dbt Fusion

Answer + AI explanation with Pro

1. A.check snapshot strategy — An incremental strategy that upserts using a unique_key, updating matched rows and inserting new ones in a single MERGE statement.
2. B.check snapshot strategy — A dbt model property (private, protected, public) that controls which projects or groups may ref a model, enforcing API boundaries in dbt Mesh.
3. C.check snapshot strategy — A snapshot strategy that detects changes by comparing a chosen set of columns for any difference, used when no reliable updated_at exists.
4. D.check snapshot strategy — A dbt layer where metrics are defined once centrally and queried consistently, so every tool computes the same numbers.

Answer + AI explanation with Pro