Databricks Certified Data Engineer Associate Exam Dumps, Practice Test Questions

100% Latest & Updated Databricks Certified Data Engineer Associate Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!

Databricks Certified Data Engineer Associate Premium Bundle
$79.97
$59.98

Certified Data Engineer Associate Premium Bundle

  • Premium File: 280 Questions & Answers. Last update: Nov 9, 2026
  • Training Course: 38 Video Lectures
  • Study Guide: 432 Pages
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates

Certified Data Engineer Associate Premium Bundle

Databricks Certified Data Engineer Associate Premium Bundle
  • Premium File: 280 Questions & Answers. Last update: Nov 9, 2026
  • Training Course: 38 Video Lectures
  • Study Guide: 432 Pages
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
$79.97
$59.98

Databricks Certified Data Engineer Associate Practice Test 17

Question 160

A telemetry source emits JSON lines with heterogeneous optional keys. The team observes: A new optional property prevents all valid records from loading. Which underlying cause most directly explains this failure?

  1. A naive batch file scan was chosen instead of incremental Auto Loader discovery.
  2. The connector destination was set to the wrong catalog or schema.
  3. A continuously active stream was chosen for an essentially one-off load.
  4. Schema handling was too strict at the raw landing boundary for this variable payload.

Explanation

Correct Answer

D:

Schema handling was too strict at the raw landing boundary for this variable payload. Ingest JSON to bronze with controlled schema handling and standardize selected fields in silver.

Incorrect Answers

A:

When a large object-store drop includes millions of files with continual arrivals, “A naive batch file scan was chosen instead of incremental Auto Loader discovery” could be relevant. Schema handling was too strict at the raw landing boundary for this variable payload.

B:

If an audit team needs stable ownership of data ingested from a managed enterprise application, “The connector destination was set to the wrong catalog or schema” may be the appropriate response. Schema handling was too strict at the raw landing boundary for this variable payload.

C:

A static set of archive files must be loaded once through SQL with minimal pipeline machinery. In those circumstances, “A continuously active stream was chosen for an essentially one-off load” fits. Schema handling was too strict at the raw landing boundary for this variable payload.

Question 161

A customer source mixes valid records with a small number of malformed JSON documents. Which design choice best addresses the stated business and technical requirement?

  1. Quarantine malformed records while allowing valid events to proceed under agreed quality rules.
  2. Use incremental file ingestion instead of rescanning an entire historical directory each run.
  3. Apply keyed upsert semantics rather than blindly appending all revisions.
  4. Choose COPY INTO for a simple repeatable bulk import.

Explanation

Correct Answer

A:

Quarantine malformed records while allowing valid events to proceed under agreed quality rules. The successful batch count plus quarantine count reconciles to input events.

Incorrect Answers

B:

A point-of-sale system publishes a new immutable object-store file every five minutes. In those circumstances, “Use incremental file ingestion instead of rescanning an entire historical directory each run” fits. Quarantine malformed records while allowing valid events to proceed under agreed quality rules.

C:

When a partner publishes corrected records with stable business keys in each incremental delivery, “Apply keyed upsert semantics rather than blindly appending all revisions” could be relevant. Quarantine malformed records while allowing valid events to proceed under agreed quality rules.

D:

If a static set of archive files must be loaded once through SQL with minimal pipeline machinery, “Choose COPY INTO for a simple repeatable bulk import” may be the appropriate response. Quarantine malformed records while allowing valid events to proceed under agreed quality rules.

Question 162

A customer source mixes valid records with a small number of malformed JSON documents. Which setup step most directly realizes the proposed design?

  1. Capture corrupt-record evidence and apply an explicit reject or quarantine path.
  2. Keep the COPY INTO target and source identity stable across reruns.
  3. Select failOnNewColumns and institute a reviewed schema-change deployment.
  4. Configure schema rescue and inspect _rescued_data for the source record.

Explanation

Correct Answer

A:

Capture corrupt-record evidence and apply an explicit reject or quarantine path. The successful batch count plus quarantine count reconciles to input events.

Incorrect Answers

B:

When a data owner receives daily Parquet exports and wants previously copied files skipped on reruns, “Keep the COPY INTO target and source identity stable across reruns” could be relevant. Capture corrupt-record evidence and apply an explicit reject or quarantine path.

C:

If a customer source cannot accept newly introduced columns without explicit schema approval, “Select failOnNewColumns and institute a reviewed schema-change deployment” may be the appropriate response. Capture corrupt-record evidence and apply an explicit reject or quarantine path.

D:

A JSON source contains additional nested keys that are not part of a strongly typed downstream contract. In those circumstances, “Configure schema rescue and inspect _rescued_data for the source record” fits. Capture corrupt-record evidence and apply an explicit reject or quarantine path.

Question 163

A customer source mixes valid records with a small number of malformed JSON documents. What is the most relevant verification evidence for this case?

  1. The successful batch count plus quarantine count reconciles to input events.
  2. The ingested tables are discoverable under the approved UC namespace and grants.
  3. Historical rows appear first and later changes are reflected without full custom reimports.
  4. A corrected key changes its existing target row without creating another active duplicate.

Explanation

Correct Answer

A:

The successful batch count plus quarantine count reconciles to input events.

Incorrect Answers

B:

If an audit team needs stable ownership of data ingested from a managed enterprise application, “The ingested tables are discoverable under the approved UC namespace and grants” may be the appropriate response. The successful batch count plus quarantine count reconciles to input events.

C:

A managed enterprise-source ingestion has an initial backfill followed by periodic change updates. In those circumstances, “Historical rows appear first and later changes are reflected without full custom reimports” fits. The successful batch count plus quarantine count reconciles to input events.

D:

When a partner publishes corrected records with stable business keys in each incremental delivery, “A corrected key changes its existing target row without creating another active duplicate” could be relevant. The successful batch count plus quarantine count reconciles to input events.

Question 164

A customer source mixes valid records with a small number of malformed JSON documents. The team observes: Malformed records disappear silently and the load reports complete success. What is the most likely mechanism behind the observed defect?

  1. The pipeline discarded parse failures without a governed error-record destination.
  2. The team used plain INSERT FROM FILES without COPY INTO's loaded-file tracking.
  3. Rescue behavior or the required rescued-data column was not configured for this source.
  4. The job never persisted the inferred schema to a stable schema location.

Explanation

Correct Answer

A:

The pipeline discarded parse failures without a governed error-record destination. Capture corrupt-record evidence and apply an explicit reject or quarantine path.

Incorrect Answers

B:

An operations team wants a SQL-only one-time load of historical CSV objects stored in an approved S3 prefix. In those circumstances, “The team used plain INSERT FROM FILES without COPY INTO's loaded-file tracking” fits. The pipeline discarded parse failures without a governed error-record destination.

C:

When a regulated feed must retain unexpected JSON attributes without changing the approved table schema, “Rescue behavior or the required rescued-data column was not configured for this source” could be relevant. The pipeline discarded parse failures without a governed error-record destination.

D:

If an Auto Loader job reads many new files and needs to persist inferred schema changes independently of checkpoints, “The job never persisted the inferred schema to a stable schema location” may be the appropriate response. The pipeline discarded parse failures without a governed error-record destination.

Question 165

A retail bronze orders table contains string totals, inconsistent country codes and missing order identifiers before a silver SLA can begin. Which approach most directly satisfies the requirement?

  1. Use UNION ALL or DataFrame unionByName rather than a relational join.
  2. Apply a left anti join or equivalent NOT EXISTS with the consent-revocation keys.
  3. Use a streaming table or incremental pipeline target rather than a manually overwritten static table.
  4. Create a silver projection that casts monetary values, normalizes country codes and rejects missing required keys without overwriting bronze.

Explanation

Correct Answer

D:

Create a silver projection that casts monetary values, normalizes country codes and rejects missing required keys without overwriting bronze. Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.

Incorrect Answers

A:

Two quarterly datasets have identical columns but their records must be vertically appended and duplicates preserved. In those circumstances, “Use UNION ALL or DataFrame unionByName rather than a relational join” fits. Create a silver projection that casts monetary values, normalizes country codes and rejects missing required keys without overwriting bronze.

B:

When a GDPR deletion feed requires excluding rows whose identifiers appear on a revoked-consent list, “Apply a left anti join or equivalent NOT EXISTS with the consent-revocation keys” could be relevant. Create a silver projection that casts monetary values, normalizes country codes and rejects missing required keys without overwriting bronze.

C:

If a downstream pipeline needs incremental consumption of a continuous events feed into a governed table, “Use a streaming table or incremental pipeline target rather than a manually overwritten static table” may be the appropriate response. Create a silver projection that casts monetary values, normalizes country codes and rejects missing required keys without overwriting bronze.

Question 166

A retail bronze orders table contains string totals, inconsistent country codes and missing order identifiers before a silver SLA can begin. Which configuration or engineering action would most directly implement the requirement?

  1. Read the bronze Delta table with PySpark, use explicit casts and normalization, quarantine invalid keys, and write a separate governed silver table.
  2. Compare runs while adjusting spark.sql.shuffle.partitions within tested bounds.
  3. Join accounts LEFT OUTER JOIN payments on account_id and preserve missing payment fields.
  4. Validate the bounded dimensions and crossJoin the two small frames.

Explanation

Correct Answer

A:

Read the bronze Delta table with PySpark, use explicit casts and normalization, quarantine invalid keys, and write a separate governed silver table. Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.

Incorrect Answers

B:

If a heavily shuffled ETL join creates thousands of tiny shuffle partitions while tasks spend time scheduling, “Compare runs while adjusting spark.sql.shuffle.partitions within tested bounds” may be the appropriate response. Read the bronze Delta table with PySpark, use explicit casts and normalization, quarantine invalid keys, and write a separate governed silver table.

C:

A subscriptions dataset must preserve every account even when no payment has arrived. In those circumstances, “Join accounts LEFT OUTER JOIN payments on account_id and preserve missing payment fields” fits. Read the bronze Delta table with PySpark, use explicit casts and normalization, quarantine invalid keys, and write a separate governed silver table.

D:

When a risk team needs all combinations of five scenario assumptions and six stress categories, “Validate the bounded dimensions and crossJoin the two small frames” could be relevant. Read the bronze Delta table with PySpark, use explicit casts and normalization, quarantine invalid keys, and write a separate governed silver table.

Question 167

A retail bronze orders table contains string totals, inconsistent country codes and missing order identifiers before a silver SLA can begin. Which test would establish that the change worked as intended?

  1. Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.
  2. Queries reflect permitted underlying table changes at query time.
  3. The published schema contains only the approved fields despite extra bronze columns.
  4. The accepted plus rejected counts reconcile to the original source batch.

Explanation

Correct Answer

A:

Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.

Incorrect Answers

B:

A finance dashboard reads a curated metric that should always reflect the latest base table without storing a separately maintained aggregate. In those circumstances, “Queries reflect permitted underlying table changes at query time” fits. Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.

C:

When an account table has three irrelevant vendor-internal columns that must not reach the published silver table, “The published schema contains only the approved fields despite extra bronze columns” could be relevant. Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.

D:

If a silver quality gate must reject records missing a transaction identifier, but analysts still need an error count, “The accepted plus rejected counts reconcile to the original source batch” may be the appropriate response. Known valid orders have typed totals and normalized country values, while rejected keyless orders remain traceable to bronze.

Question 168

A retail bronze orders table contains string totals, inconsistent country codes and missing order identifiers before a silver SLA can begin. The team observes: A finance aggregate reports lexically sorted amounts and blank business identifiers after the cleansing run. Which configuration or design gap best explains the symptom?

  1. The job insisted on a full exact DISTINCT despite permitted approximate accuracy.
  2. The denominator included rows without a valid numeric price.
  3. The pipeline left the legacy column name or dropped it without a replacement.
  4. The pipeline wrote raw strings and permitted null business keys into silver without validation.

Explanation

Correct Answer

D:

The pipeline wrote raw strings and permitted null business keys into silver without validation. Read the bronze Delta table with PySpark, use explicit casts and normalization, quarantine invalid keys, and write a separate governed silver table.

Incorrect Answers

A:

A reporting table must count unique devices across billions of events with an acceptable approximation. In those circumstances, “The job insisted on a full exact DISTINCT despite permitted approximate accuracy” fits. The pipeline wrote raw strings and permitted null business keys into silver without validation.

B:

When a sales dashboard needs mean transaction value per region while excluding null prices, “The denominator included rows without a valid numeric price” could be relevant. The pipeline wrote raw strings and permitted null business keys into silver without validation.

C:

If a customer schema is changing from old_customer_id to customer_id while dashboards expect the new name, “The pipeline left the legacy column name or dropped it without a replacement” may be the appropriate response. The pipeline wrote raw strings and permitted null business keys into silver without validation.

Question 169

An IoT bronze dataset contains temperatures as strings and an impossible reading that must not contaminate hourly gold summaries. Which architecture or engineering decision best fits this scenario?

  1. Use UNION ALL or DataFrame unionByName rather than a relational join.
  2. Split the field into separate city and postal-code columns with a validated schema.
  3. Enforce numeric conversion and an allowed sensor range during the bronze-to-silver transform.
  4. Tune shuffle partition count using observed data volume and task metrics rather than a fixed guess.

Explanation

Correct Answer

C:

Enforce numeric conversion and an allowed sensor range during the bronze-to-silver transform. The impossible reading is segregated, valid floats remain queryable, and counts reconcile with bronze.

Incorrect Answers

A:

When two quarterly datasets have identical columns but their records must be vertically appended and duplicates preserved, “Use UNION ALL or DataFrame unionByName rather than a relational join” could be relevant. Enforce numeric conversion and an allowed sensor range during the bronze-to-silver transform.

B:

If a source field packs a city and postal code with a single pipe delimiter, “Split the field into separate city and postal-code columns with a validated schema” may be the appropriate response. Enforce numeric conversion and an allowed sensor range during the bronze-to-silver transform.

D:

When a heavily shuffled ETL join creates thousands of tiny shuffle partitions while tasks spend time scheduling, “Tune shuffle partition count using observed data volume and task metrics rather than a fixed guess” could be relevant. Enforce numeric conversion and an allowed sensor range during the bronze-to-silver transform.

Question 170

An IoT bronze dataset contains temperatures as strings and an impossible reading that must not contaminate hourly gold summaries. Which implementation step is most appropriate before production use?

  1. Use try_cast or a controlled parser, range predicates, and a quarantined invalid-record output before saving silver.
  2. Create the appropriate Unity Catalog view with an explicit stable SELECT.
  3. Group the relevant rows and apply approx_count_distinct to device_id.
  4. Define a supported materialized view in Unity Catalog and configure refresh according to requirements.

Explanation

Correct Answer

A:

Use try_cast or a controlled parser, range predicates, and a quarantined invalid-record output before saving silver. The impossible reading is segregated, valid floats remain queryable, and counts reconcile with bronze.

Incorrect Answers

B:

A finance dashboard reads a curated metric that should always reflect the latest base table without storing a separately maintained aggregate. In those circumstances, “Create the appropriate Unity Catalog view with an explicit stable SELECT” fits. Use try_cast or a controlled parser, range predicates, and a quarantined invalid-record output before saving silver.

C:

When a reporting table must count unique devices across billions of events with an acceptable approximation, “Group the relevant rows and apply approx_count_distinct to device_id” could be relevant. Use try_cast or a controlled parser, range predicates, and a quarantined invalid-record output before saving silver.

D:

If a busy BI dashboard repeatedly runs an expensive grouped query and tolerates refresh-based staleness, “Define a supported materialized view in Unity Catalog and configure refresh according to requirements” may be the appropriate response. Use try_cast or a controlled parser, range predicates, and a quarantined invalid-record output before saving silver.

Certified Data Engineer Associate Practice Tests

UP

SPECIAL OFFER: GET 10% OFF

This is ONE TIME OFFER

ExamSnap Discount Offer
Enter Your Email Address to Receive Your 10% Off Discount Code

A confirmation link will be sent to this email address to verify your login. *We value your privacy. We will not rent or sell your email address.

Download Free Demo of VCE Exam Simulator

Experience Avanset VCE Exam Simulator for yourself.

Simply submit your e-mail address below to get started with our interactive software demo of your free trial.

Free Demo Limits: In the demo version you will be able to access only first 5 questions from exam.