I85PREP
⌘ K
Cloud Data Quality R41 Professional study materials
Cloud Data Quality R41 Professional study resources

Cloud Data Quality R41 Professional practice questions

One question at a time. Build your Cloud Data Quality R41 Professional understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Cloud Data Quality

In a cloud data quality pipeline, a rule that flags records where the value in a mandatory field is NULL is best classified as which type of check?

  1. Referential integrity check
  2. Format conformity check
  3. Completeness check Correct answer.
  4. Uniqueness check

Explanation

Completeness measures whether required values are present, so a NULL test on a mandatory field is a completeness rule. Uniqueness, referential integrity, and format checks each target different properties of the data.

Question 2Cloud Data Quality

A data quality team wants to detect duplicate customer records caused by the same person being entered with slightly different spellings. Which technique is most appropriate?

  1. Row-count reconciliation between source and target
  2. Column-level null profiling
  3. Exact-match primary key comparison
  4. Fuzzy matching on name and address attributes Correct answer.

Explanation

Near-duplicate detection across variations in spelling requires fuzzy or probabilistic matching rather than exact equality. Exact-match comparison only catches identical values, and the other techniques do not address similarity.

Question 3Cloud Data Quality

Which cloud pattern most directly supports near-real-time data quality monitoring on a stream of events?

  1. Quarterly data steward review meetings
  2. Nightly batch profiling of the landing table
  3. Manual sampling of records once per release
  4. Streaming rules evaluated on each event or micro-batch Correct answer.

Explanation

Near-real-time quality monitoring requires rules to be evaluated as events arrive, typically in a streaming or micro-batch engine. Batch, manual sampling, and periodic review all introduce latency.

Question 4Cloud Data Quality

A quality scorecard shows that a column is 99.9% complete but the business still reports missing customer records. Which interpretation is most sound?

  1. The business complaint must be incorrect
  2. Completeness is always an unreliable dimension
  3. The completeness rule is measuring the wrong column or scope Correct answer.
  4. The scorecard should be discarded

Explanation

A high completeness score alongside business complaints usually indicates the rule does not cover the column or population that matters. The correct response is to re-scope the rule, not to abandon the metric.

Question 5Cloud Data Quality

Which approach best implements a quarantine pattern for records that fail data quality rules in a cloud pipeline?

  1. Silently dropping failing records
  2. Routing failing records to a separate table with the failure reason Correct answer.
  3. Overwriting the source system with corrected values
  4. Halting the entire pipeline on the first failure

Explanation

Quarantine preserves failing records in a separate store together with the reason for failure so they can be remediated and audited. Silently dropping them loses evidence, and halting the pipeline blocks all processing.

Question 6Cloud Data Quality

A rule flags order records whose ship date precedes the order date. Which data quality dimension does this rule primarily test?

  1. Timeliness
  2. Accuracy
  3. Validity
  4. Consistency Correct answer.

Explanation

Cross-field logical rules such as ship date not preceding order date test internal consistency of the record. Timeliness concerns latency, accuracy concerns correspondence to reality, and validity concerns conformance to a defined domain.

Question 7Cloud Data Quality

Which metadata is most important to record alongside each data quality rule result to support later root-cause analysis?

  1. The rule definition, execution timestamp, dataset version, and failing record count Correct answer.
  2. The dashboard color theme
  3. The rule author's department name only
  4. The tool's license expiry date

Explanation

Root-cause analysis needs the rule logic, when it ran, which data version was assessed, and how many records failed. This context lets analysts reproduce and trace failures over time.

Question 8Cloud Data Quality

A team wants to enforce a rule that every value in a currency column belongs to a fixed list of ISO 4217 codes. Which check is this?

  1. Domain or enumerated-value check Correct answer.
  2. Range check
  3. Cross-field consistency check
  4. Uniqueness check

Explanation

Restricting a column to a fixed list of allowed values is a domain or enumerated-value validity check. Range checks constrain numeric intervals, uniqueness forbids duplicates, and cross-field checks compare multiple fields.

Question 9Cloud Data Quality

Which metric best communicates the share of records that passed all applicable data quality rules during a run?

  1. Rule pass rate Correct answer.
  2. Query latency
  3. Average column width
  4. Total row count

Explanation

Rule pass rate expresses the proportion of records satisfying the rules and is a direct quality indicator. Row count, column width, and latency describe volume or performance rather than quality.

Question 10Cloud Data Quality

A cloud data platform ingests data from three source systems that use different country code formats. Which design most effectively enforces consistent country values downstream?

  1. Allowing each consumer to normalize codes independently
  2. Standardizing codes to ISO 3166 in a curated layer before consumption Correct answer.
  3. Removing the country column from the curated layer
  4. Storing the raw formats unchanged in the reporting table

Explanation

Central standardization in a curated layer ensures all consumers see a single canonical representation. Pushing normalization to each consumer multiplies effort and risk of divergence.

Question 11Cloud Data Quality

Which statement best describes the role of a data contract between a producing team and consuming teams in a cloud environment?

  1. It is a dashboard that displays row counts
  2. It is a legal document that transfers ownership of data
  3. It defines agreed schema, semantics, and quality expectations enforced at the producer boundary Correct answer.
  4. It replaces the need for any data quality testing

Explanation

A data contract codifies schema, meaning, and quality expectations so producers deliver data that meets consumer needs. It complements testing rather than replacing it and is not a legal ownership transfer.

Question 12Cloud Data Quality

A referential integrity rule on a fact table's customer key reports many orphans after a nightly load. Which cause is most likely?

  1. The database storage is compressed
  2. The dimension load ran after the fact load Correct answer.
  3. The fact table has too many indexes
  4. The reporting layer uses a different color palette

Explanation

Orphan foreign keys commonly occur when facts are loaded before the referenced dimension rows exist. Sequencing the dimension load first, or deferring enforcement, resolves the timing issue.

Question 13Cloud Data Quality

Which technique best detects drift in the distribution of a numeric column across successive loads?

  1. Comparing summary statistics such as mean, min, max, and percentiles over time Correct answer.
  2. Counting the number of columns in the table
  3. Listing the table's owners
  4. Reviewing the SQL dialect used by the warehouse

Explanation

Distribution drift is detected by tracking summary statistics and comparing them across time windows. Schema counts, ownership, and dialect do not describe value distributions.

Question 14Cloud Data Quality

A data quality rule is defined as 'email must match a valid pattern'. Which implementation concern matters most for avoiding false positives?

  1. Choosing a pattern that accepts all valid address forms used by the business Correct answer.
  2. Running the rule once per year
  3. Making the regex as short as possible
  4. Applying the rule only to null values

Explanation

Overly strict patterns reject legitimate addresses, producing false positives that erode trust in the rule. The pattern must reflect the address forms the business actually accepts.

Question 15Cloud Data Quality

Which cloud practice most improves reproducibility of data quality test outcomes across environments?

  1. Version-controlling rule definitions and test data alongside pipeline code Correct answer.
  2. Storing rule logic only in a wiki page
  3. Relying on tribal knowledge of the original author
  4. Allowing analysts to edit rules directly in production

Explanation

Version-controlled rule definitions and fixtures make test outcomes reproducible and reviewable. Ad hoc production edits and undocumented knowledge create inconsistency.

Question 16Cloud Data Quality

Which statement correctly distinguishes data quality rules from data quality SLAs?

  1. Rules are legal contracts; SLAs are SQL scripts
  2. Rules define measurable conditions; SLAs define the acceptable thresholds and consequences Correct answer.
  3. Rules apply only to streaming data; SLAs apply only to batch
  4. Rules and SLAs are synonyms

Explanation

Rules specify the condition being tested, while SLAs set the target level of performance or quality and the response when it is missed. They are complementary but distinct concepts.

Question 17Cloud Data Quality

A pipeline applies a uniqueness rule to a column that legitimately contains repeated values for order line items. What is the most appropriate correction?

  1. Apply uniqueness to the composite key of order and line number Correct answer.
  2. Move the column to a different table
  3. Add a random suffix to duplicate values
  4. Remove the uniqueness rule and accept duplicates

Explanation

The uniqueness rule was scoped to the wrong grain; line items are unique only within an order. Testing the composite key reflects the true business key.

Question 18Cloud Data Quality

Which design best prevents a quality rule from being silently disabled during a schema change?

  1. Treating missing referenced columns as a rule failure that blocks promotion Correct answer.
  2. Running rules only on demand
  3. Documenting column names in a spreadsheet
  4. Wrapping rule execution in a try/except that ignores all errors

Explanation

Failing closed when referenced columns disappear ensures schema changes surface as explicit failures rather than silent gaps. Swallowing exceptions hides the problem.

Question 19Cloud Data Quality

A team observes that a rule's failure count spikes only on weekends. Which investigation step is most appropriate first?

  1. Disable the rule on weekends
  2. Assume the data is correct and ignore the spike
  3. Increase the threshold until the spike disappears
  4. Compare source feed schedules and volumes across weekdays and weekends Correct answer.

Explanation

A pattern tied to specific days suggests a change in upstream feed behavior, such as weekend batch timing or volume. Investigating the source schedule identifies the cause before any rule change.

Question 20Cloud Data Quality

Which approach best quantifies the business impact of a data quality defect for prioritization?

  1. Counting the number of columns in the schema
  2. Measuring the size of the rule definition file
  3. Listing the number of rules in the catalog
  4. Counting the number of affected records and mapping them to downstream reports or decisions Correct answer.

Explanation

Impact depends on how many records are affected and which business processes consume them. Linking defects to downstream use lets teams rank remediation by consequence.

Question 21Cloud Data Quality

A cloud data catalog automatically profiles new tables. Which output most directly helps stewards decide where to invest quality rules?

  1. The IAM role of the loader
  2. Column-level null and distinct-value statistics Correct answer.
  3. The warehouse compute tier
  4. The physical storage region

Explanation

Null rates and distinct-value counts reveal which columns are sparse, constant, or highly variable, guiding rule selection. Infrastructure attributes do not describe data content.

Question 22Cloud Data Quality

Which scenario best illustrates a timeliness defect rather than an accuracy defect?

  1. The dashboard shows the wrong product price
  2. The dashboard shows a duplicated order
  3. The dashboard shows yesterday's inventory instead of today's Correct answer.
  4. The dashboard shows a malformed date

Explanation

Timeliness concerns whether data is current enough for its purpose; stale inventory is a latency issue. Wrong prices, duplicates, and malformed dates concern accuracy, uniqueness, and validity.

Question 23Cloud Data Quality

What is the primary benefit of defining data quality rules as code stored in a repository rather than configuring them manually in a UI?

  1. Rules can be reviewed, tested, and promoted through the same pipeline as application code Correct answer.
  2. Rules never need to be changed
  3. Rules no longer require any data
  4. Rules automatically fix all data defects

Explanation

Rules as code enable peer review, automated testing, and consistent promotion across environments. They still require maintenance and do not automatically remediate data.

Question 24Cloud Data Quality

Which control best limits the blast radius of a faulty data quality rule that incorrectly quarantines most of a day's records?

  1. A staged rollout with sampled shadow evaluation before full enforcement Correct answer.
  2. Skipping review because the rule is new
  3. Immediate full enforcement in production
  4. Removing the quarantine table

Explanation

Evaluating a new rule in shadow mode on samples reveals false positives before it affects production volumes. Full enforcement without validation risks large-scale misclassification.

Question 25Cloud Data Quality

A quality dashboard reports 100% conformance for a rule that has never failed since deployment. Which explanation warrants the most scrutiny?

  1. The rule should be deleted immediately
  2. The dashboard is always correct
  3. The data is guaranteed perfect
  4. The rule may be misconfigured or evaluating an empty or wrong dataset Correct answer.

Explanation

A rule that never fails is suspicious and may be misconfigured, scoped to the wrong table, or running against no rows. Investigating configuration is prudent before trusting the result.

Question 26Cloud Data Quality

A data engineer needs a rule that checks every row of a transactions table for a non-null account_id before the table is promoted to the curated zone. Which rule type is this?

  1. Completeness rule Correct answer.
  2. Uniqueness rule
  3. Range rule
  4. Referential integrity rule

Explanation

Completeness rules verify that required attributes are populated, so a non-null check on account_id is a completeness assertion. Uniqueness, referential integrity, and range rules test different properties and would not catch a missing value.

Question 27Cloud Data Quality

Which data quality dimension is measured when a rule verifies that a customer's birth_date is never later than the current date?

  1. Timeliness
  2. Consistency
  3. Validity Correct answer.
  4. Completeness

Explanation

Validity checks whether values conform to defined business or format constraints, and a future birth date violates a domain constraint. Timeliness concerns latency of delivery, not plausibility of values.

Question 28Cloud Data Quality

A steward wants a single number that shows the percentage of rows satisfying a specific rule in the latest run. Which metric should be reported?

  1. Rule pass rate Correct answer.
  2. Data freshness lag
  3. Schema drift count
  4. Quarantine volume

Explanation

A rule pass rate expresses the proportion of evaluated rows that satisfied the rule, which is exactly the requested percentage. Freshness lag, drift count, and quarantine volume measure other aspects of the pipeline.

Question 29Cloud Data Quality

Which artifact should store the expected schema, allowed value sets, and nullability constraints that consuming teams rely on from a producer?

  1. A runbook
  2. A data contract Correct answer.
  3. A lineage graph
  4. A cost report

Explanation

A data contract codifies the schema, semantics, and quality expectations a producer guarantees to consumers. Runbooks, lineage graphs, and cost reports document operations or dependencies rather than guarantees.

Question 30Cloud Data Quality

Which practice makes data quality test results comparable when the same suite runs in development, staging, and production?

  1. Allowing each environment to tune thresholds independently without recording changes
  2. Running the suite only in production to avoid environment noise
  3. Relying on analysts to apply rules manually in each environment
  4. Pinning rule definitions and thresholds in version control and deploying them identically Correct answer.

Explanation

Version-controlled rule definitions deployed identically ensure that differences in results reflect data, not configuration drift. Independent tuning or manual application makes cross-environment comparison meaningless.

Question 31Cloud Data Quality

A rule enforces that each order_id appears exactly once in an orders table. Which data quality dimension does this rule directly test?

  1. Validity
  2. Uniqueness Correct answer.
  3. Accuracy
  4. Timeliness

Explanation

Uniqueness asserts that a key value occurs only once, which is precisely what an order_id duplicate check verifies. Accuracy, timeliness, and validity address correctness against reality, delivery latency, and format conformance respectively.

Question 32Cloud Data Quality

Which output from an automated data profiling job most directly helps a steward choose which columns need quality rules first?

  1. Query execution plans for the profiling job
  2. Null ratio and distinct-value counts per column Correct answer.
  3. Storage cost per table partition
  4. Number of downstream dashboards per table

Explanation

Null ratios and distinct-value counts reveal where completeness and cardinality problems are concentrated, guiding rule prioritization. Execution plans, storage cost, and dashboard counts inform performance or cost decisions rather than data quality risk.

Question 33Cloud Data Quality

A nightly job loads records into a reporting table at 03:00, but the business requires the table refreshed by 01:00. Which defect does this represent?

  1. A completeness defect
  2. A uniqueness defect
  3. An accuracy defect
  4. A timeliness defect Correct answer.

Explanation

Timeliness defects occur when data is not available within the required window even if its values are correct. Accuracy, uniqueness, and completeness describe the content of the data rather than when it becomes available.

Question 34Cloud Data Quality

A rule that checks order totals against line-item sums reports failures only for a small set of records. Which metadata should be attached to each failure to support root-cause analysis?

  1. The rule version, source table, load batch identifier, and execution timestamp Correct answer.
  2. Only the rule owner's name
  3. Only the total failure count for the run
  4. Only the dashboard that displays the result

Explanation

Capturing rule version, source, batch, and timestamp lets analysts correlate failures with specific loads and rule changes. A bare count or owner name cannot localize the defect to a batch or a rule revision.

Question 35Cloud Data Quality

A currency column must contain only codes from a fixed ISO 4217 list. Which check best enforces this?

  1. A row-count reconciliation check
  2. An allowed-values or domain-membership check Correct answer.
  3. A regex check for three uppercase letters
  4. A non-null check

Explanation

A domain-membership check compares each value against the authoritative code list, which is the only option that rejects invalid but well-formed codes. A pattern check accepts any three uppercase letters, including codes that do not exist.

Question 36Cloud Data Quality

A fact table's customer_key shows many orphans after a nightly load. Which cause is most likely?

  1. The reporting layer applied a filter to exclude inactive customers
  2. The fact table's partitioning scheme changed between runs
  3. The fact table's numeric measures were aggregated incorrectly
  4. The dimension load completed after the fact load within the same window Correct answer.

Explanation

Orphan keys typically appear when facts arrive before their matching dimension rows exist, which is a sequencing problem. Measure aggregation, report filters, and partitioning affect results but do not create missing parent keys.

Question 37Cloud Data Quality

A team wants to detect gradual drift in the distribution of a numeric amount column across successive loads. Which technique is most appropriate?

  1. Validating the column's declared data type
  2. Checking only that the column is non-null in each load
  3. Comparing per-load summary statistics such as mean, percentiles, and histogram shape against a baseline Correct answer.
  4. Counting the number of distinct values in the latest load

Explanation

Distribution drift is detected by comparing statistical summaries and histogram shape across loads against an established baseline. Null checks, distinct counts, and type validation catch different problems and would miss a slow shift in the distribution.

Question 38Cloud Data Quality

A regex-based email rule is being deployed. Which implementation concern matters most for avoiding false failures?

  1. Running the rule only on records inserted in the last hour
  2. Applying the pattern only to the local part of the address
  3. Making the pattern case-sensitive to preserve data fidelity
  4. Anchoring the pattern to the whole value and testing it against a representative sample of real addresses Correct answer.

Explanation

Unanchored patterns can match substrings and produce incorrect results, and validating against real samples exposes false positives before deployment. Partial matching, forced case sensitivity, and narrow scoping all increase the risk of spurious failures.

Question 39Cloud Data Quality

A rule's failure count spikes sharply only on weekend loads. Which investigation step is most appropriate first?

  1. Escalate to the vendor without gathering evidence
  2. Disable the rule until the following month
  3. Compare the weekend load's source extract, schedule, and upstream job versions with weekday runs Correct answer.
  4. Immediately tighten the rule threshold to suppress the spike

Explanation

Isolating what differs about weekend runs, such as source extracts, schedules, or job versions, identifies the cause before any remediation. Suppressing, disabling, or escalating prematurely hides the defect or wastes effort.

Question 40Cloud Data Quality

A quality dashboard shows 100% conformance for a rule that has never failed since it was deployed. Which explanation warrants investigation first?

  1. The rule's owner has changed teams recently
  2. The dashboard is aggregating results from multiple environments
  3. The rule is performing exactly as designed and needs no review
  4. The rule may be scoped to a filter that matches no rows or an outdated column Correct answer.

Explanation

A rule that never fails is suspicious when the underlying data is expected to vary, and an empty scope or stale column reference would produce vacuous passes. Assuming correct behavior, aggregation, or ownership changes does not explain the absence of failures.

Question 41Cloud Data Quality

A team must choose between configuring quality rules in a vendor UI and defining them as code in a repository. Which benefit most strongly favors rules as code?

  1. Faster rule authoring for non-technical stewards
  2. Peer review, version history, and repeatable promotion across environments Correct answer.
  3. Elimination of the need for test data
  4. Automatic resolution of data defects

Explanation

Rules as code enable code review, change history, and consistent promotion through environments, which improves governance and reproducibility. It does not eliminate test data needs or automatically fix defects, and it is typically less approachable for non-technical users.

Question 42Cloud Data Quality

A faulty rule begins quarantining most of a day's records. Which control best limits the blast radius?

  1. Increasing the quarantine table's retention period
  2. A canary or sampled dry run that reports impact before full enforcement Correct answer.
  3. Adding more dashboards to monitor the rule
  4. Granting the rule broader write permissions

Explanation

A sampled dry run measures how many records a rule would reject before it is enforced, allowing thresholds and logic to be corrected first. Retention, dashboards, and permissions affect visibility or storage but not the scale of the impact.

Question 43Cloud Data Quality

A pipeline ingests country values from three systems using ISO alpha-2, alpha-3, and full names. Which design most sustainably standardizes these values for downstream quality rules?

  1. A separate rule for each source format that flags non-conforming values
  2. A governed reference mapping applied during ingestion that converts each format to a canonical code Correct answer.
  3. A free-text field retained as-is with rules applied per report
  4. Manual correction by analysts after each load

Explanation

A governed reference mapping normalizes heterogeneous formats to one canonical representation, so downstream rules test a single domain. Per-source flags, free text, and manual fixes leave inconsistency in place and scale poorly.

Question 44Cloud Data Quality

A uniqueness rule is applied to a column that legitimately repeats values for order line items. What is the most appropriate correction?

  1. Redefine the rule to test the composite key of order_id and line_number Correct answer.
  2. Keep the rule and accept the failures as expected noise
  3. Change the rule to a non-null check on the same column
  4. Remove the column from the table

Explanation

The rule's grain was wrong: uniqueness should be asserted on the key that actually identifies a row, which is the composite of order_id and line_number. Accepting noise or removing the column does not fix the mis-specified assertion.

Question 45Cloud Data Quality

A team wants to ensure a quality rule is not silently disabled when a source schema changes. Which design best achieves this?

  1. Log a warning when a referenced column disappears and continue processing
  2. Store rules in a spreadsheet reviewed annually
  3. Fail the pipeline when a rule's referenced columns are missing and require explicit rule review on schema change Correct answer.
  4. Allow rules to skip evaluation when columns are absent

Explanation

Treating missing referenced columns as a hard failure forces an explicit decision rather than allowing the rule to lapse. Warnings, annual reviews, and silent skipping all permit undetected loss of coverage.

Question 46Cloud Data Quality

A team must prioritize remediation among several quality defects. Which approach best quantifies business impact?

  1. Rank defects by the rule owner's seniority
  2. Rank defects by the order in which they were discovered
  3. Rank defects by the number of failed rows only
  4. Estimate affected downstream reports, decisions, and revenue exposure per defect Correct answer.

Explanation

Business impact is quantified by tracing defects to the reports, decisions, and revenue they influence, which supports prioritization beyond raw counts. Row counts, discovery order, and ownership do not capture downstream consequence.

Question 47Cloud Data Quality

A rule flags order records whose ship_date precedes the order_date. Which data quality dimension does this rule primarily test?

  1. Completeness
  2. Timeliness
  3. Consistency Correct answer.
  4. Uniqueness

Explanation

Cross-field logic that must hold between two related dates is a consistency assertion about internal agreement of the record. Completeness, uniqueness, and timeliness address missing values, duplicates, and delivery latency rather than logical contradiction.

Question 48Cloud Data Quality

Which statement correctly distinguishes a data quality rule from a data quality SLA?

  1. A rule applies only to production, while an SLA applies only to development
  2. A rule is a contractual commitment, while an SLA is a technical assertion
  3. A rule defines the assertion evaluated against data, while an SLA sets the target and response commitment for results Correct answer.
  4. A rule is owned by the vendor, while an SLA is owned by the pipeline

Explanation

Rules express the test itself, whereas SLAs define the acceptable level of conformance and the consequences or response when it is missed. Reversing the definitions or assigning environment-specific and ownership-based scopes misstates both concepts.

Question 49Cloud Data Quality

A quality rule passes every run, but downstream analysts still report bad data. Which explanation is most plausible?

  1. The rule engine is executing too frequently
  2. The rule thresholds were set by the data steward
  3. The rule set covers only a subset of the columns and constraints that analysts actually depend on Correct answer.
  4. The results are stored in a cloud warehouse

Explanation

A passing suite proves only that the asserted rules hold; if the assertions do not cover the columns and constraints analysts rely on, defects pass through undetected. Execution frequency, threshold ownership, and storage location do not explain unreported defects.

Question 50Cloud Data Quality

A referential integrity rule on a fact table's product_key begins failing after a new source system is onboarded. Which analysis path is most rigorous?

  1. Exclude the new source's records from the fact table
  2. Lower the rule's severity until the failures stop
  3. Profile orphan keys against the dimension's load timing, key derivation logic, and source-to-target mapping for the new system Correct answer.
  4. Add the orphan keys to the dimension without review

Explanation

Orphans introduced by a new source usually trace to timing, key derivation, or mapping differences, so profiling against those factors isolates the cause. Masking severity, blind key insertion, or excluding records conceals the defect and corrupts the model.