Implementing Data Engineering Solutions Using Azure Databricks 온라인 연습
최종 업데이트 시간: 2026년06월04일
당신은 온라인 연습 문제를 통해 Microsoft DP-750 시험지식에 대해 자신이 어떻게 알고 있는지 파악한 후 시험 참가 신청 여부를 결정할 수 있다.
시험을 100% 합격하고 시험 준비 시간을 35% 절약하기를 바라며 DP-750 덤프 (최신 실제 시험 문제)를 사용 선택하여 현재 최신 100개의 시험 문제와 답을 포함하십시오.
/ 6
Question No : 1
A data engineer maintains a derived reporting table used by dashboards. The table must be rebuilt from a SELECT query during deployment, but downstream users should continue referencing the same table name.
Which statement is most appropriate?
정답: Explanation:
CREATE OR REPLACE TABLE AS SELECT rebuilds a table from a query result while preserving the object name used by downstream users.
Question No : 2
After identifying a bad write to a Delta table, the team wants to revert the table to a previous known-good version.
Which operation should be used?
정답: Explanation:
RESTORE can revert a Delta table to a previous version or timestamp when the required files are still available.
Question No : 3
A batch process updates a Delta table incorrectly. The team needs to identify which operation introduced the bad data before deciding how to recover.
Which command or capability should be used first?
정답: Explanation:
DESCRIBE HISTORY shows Delta table versions and operations, helping identify when a problematic write occurred.
Question No : 4
A source JSON field named amount sometimes arrives as an integer and sometimes as a string. The silver table requires a decimal type for accurate calculations.
What should the transformation logic do?
정답: Explanation:
The transformation should cast the field to the required numeric type and validate invalid values before promotion.
Question No : 5
A pipeline must prevent negative transaction amounts from being promoted into curated tables. The rule should be enforced as part of the data quality design rather than detected only after reporting.
Which two controls are most relevant? Choose two.
정답: Explanation:
A CHECK constraint can enforce valid values at the table level, while a data quality expectation can enforce or track quality rules in a pipeline.
Question No : 6
A downstream process must consume only rows that changed in a Delta table since the previous run, including inserts, updates, and deletes.
Which Delta Lake feature should be considered?
정답: Explanation:
Change Data Feed allows downstream processes to read row-level changes from a Delta table.
Question No : 7
A data engineer accidentally configures two independent Structured Streaming queries to use the same checkpoint directory.
What is the most likely issue?
정답: Explanation:
Each streaming query should use a separate checkpoint location to avoid conflicts in progress and state tracking.
Question No : 8
A Delta table has obsolete files from previous versions. The team wants to reduce storage cost but must preserve the required time travel retention period.
Which operation should be used carefully?
정답: Explanation:
VACUUM removes obsolete files after the retention period, but aggressive use can affect time travel and restore.
Question No : 9
A Delta table receives frequent streaming writes that create many small files. Query performance is degrading because queries scan too many files.
Which operation should be scheduled?
정답: Explanation:
OPTIMIZE compacts small files into larger files, which can improve query performance.
Question No : 10
A company wants to publish curated business metrics for executive dashboards. The data should be cleaned, aggregated, and optimized for consumption.
Which medallion layer is most appropriate?
정답: Explanation:
The gold layer is typically used for curated, business-ready data products and aggregated metrics.
Question No : 11
A team is building a bronze-silver-gold architecture. The bronze layer must support replay, auditing, and troubleshooting of source data issues.
Which design is most appropriate?
정답: Explanation:
The bronze layer commonly stores raw or minimally transformed data to support replay and troubleshooting.
Question No : 12
A pipeline reads CSV files from cloud storage. Some rows contain unexpected extra columns or malformed values. The engineering team wants to retain problematic data for later review instead of silently losing it.
Which capability is most relevant?
정답: Explanation:
Rescued data handling helps retain fields or values that do not match the expected schema for later inspection.
Question No : 13
A data engineer compares yesterday's customer snapshot with today's customer snapshot. The goal is to identify customer records that existed yesterday but are missing today, without combining both datasets.
Which operation best fits the requirement?
정답: Explanation:
EXCEPT or a left anti join can return rows from yesterday's snapshot that are not present in today's snapshot.
Question No : 14
A Lakeflow Spark Declarative Pipeline loads order data into a silver table. If order_id is null, the record must not be promoted, and the quality violation must be visible to operators.
Which feature should be used?
정답: Explanation:
Pipeline expectations define data quality rules and can fail, drop, or track invalid records during pipeline execution.
Question No : 15
A customer dimension must preserve every status change. Analysts need to query both the current status and the status that was valid on a previous date.
Which design should be implemented?
정답: Explanation:
SCD Type 2 stores historical versions of dimension records, usually with effective dates and a current indicator.
