Question 600 An ingestion job appends every day to a bronze table but must prevent an internal reader from mistaking incomplete daily input for finalized reporting data. What is the safer publish contract? Allow gold reports to assume every append event means all supplier files arrived Write a completion/control record and promote only reconciled daily loads to trusted consumers Have downstream reports read the bronze daily partition as soon as ingestion begins Assume a scheduled daily batch is complete as soon as the append job starts Explanation Correct Answer B:Explicit batch completeness evidence separates partial ingestion from accepted reporting snapshots. Incorrect Answers A:A write event cannot prove all expected deliveries are complete. C:A live bronze partition can be incomplete; the consumer needs a tested completion and reconciliation boundary before treating it as final. D:A started append is not a reconciled publication boundary; downstream readers can see partial or unapproved daily data.
Question 601 A cloud object store delivers many files with a `.json.gz` suffix. Which reader behavior should be checked before designing a custom decompression step? Replace all files with notebook outputs Whether the configured input format supports the source compression codec transparently Assume every gzip file must be copied to a public bucket first Change the logical schema to one Boolean Explanation Correct Answer B:Supported Spark readers can often decode recognized compressed inputs without a manual pre-decompression workflow. Incorrect Answers A:Notebook output is not a general substitute for reading compressed source files. C:Public copying is unnecessary and creates a security risk. D:Compression format does not imply different business field types.
Question 602 A source uses a fixed-length binary identifier containing arbitrary bytes. An engineer proposes decoding it as UTF-8 text without checking validity. What risk arises? UTF-8 decoding always preserves every arbitrary byte sequence Binary bytes may not map to valid text and the identifier can be corrupted Delta requires all identifiers to be floating-point numbers Decoding binary changes only table column comments Explanation Correct Answer B:The source data contract should determine binary or encoded representation before conversion. Incorrect Answers A:Not every byte sequence is valid UTF-8 text. C:Delta supports typed values and does not force binary identity into numeric representation. D:The stored values can change or error during inappropriate decoding.
Question 603 An upstream team changes a numeric field from integer to decimal during a quarterly feed update. What should ingestion validate? Only whether the filename still ends in CSV Assume any integer-to-decimal widening is safe without checking downstream precision Backward-compatible type widening and downstream precision requirements Drop decimal places without notifying downstream consumers Explanation Correct Answer C:A proposed type change must be assessed for schema support and semantic precision before promotion. Incorrect Answers A:Identical file suffixes can contain changed field types. B:A widening transformation may preserve range yet change expected precision or consumer contracts, so compatibility requires validation. D:Silent truncation can corrupt financial values.
Question 604 A JSON source suddenly adds a `customer_attributes` object with personally identifying fields. The existing schema evolution mechanism can ingest it. What additional decision is necessary? Determine classification, access control and accepted schema contract before exposing the new field Assume automatically evolved fields are always approved for every user Ingest the field as ordinary data and grant unchanged downstream access automatically Skip the new field permanently without retaining raw provenance Explanation Correct Answer A:Technically accepting a new field does not establish that downstream users are permitted to read its sensitive values. Incorrect Answers B:Schema detection is not privacy review. C:Automatic schema evolution can expose sensitive new fields unless classification and privileges are reviewed before publication. D:Silently losing the new source field removes evidence and may violate the ingest contract.
Question 605 A vendor sends events with UTC timestamps but also a local timezone offset field. Analysts need local business-day grouping later. What is the right bronze storage posture? Replace the instant with ingestion date Drop timezone details because all local dates equal UTC dates Preserve the authoritative UTC instant and original offset or zone context Convert timestamps to an unvalidated server locale Explanation Correct Answer C:Keeping the instant and timezone context supports correct later localization without lossy assumptions. Incorrect Answers A:Ingestion date is not the event occurrence time. B:Local business dates can differ from UTC near midnight. D:Server locale may not match the source's business timezone.
Question 606 A feed includes local timestamps during a daylight-saving fall-back hour when 01:30 occurs twice. What makes the source event time unambiguous? Only the ingestion job task ID Lexicographic sort of the local timestamp alone Round every timestamp to midnight An explicit UTC offset, time zone with disambiguation, or authoritative UTC instant Explanation Correct Answer D:An ambiguous local wall-clock timestamp cannot always be mapped to one instant without additional context. Incorrect Answers A:Task IDs do not identify source time ambiguity. B:The two events can have the same local time text but different instants. C:Rounding destroys the distinction rather than resolving it.
Question 607 An ingestion pipeline receives an API array containing optional fields that are absent in some objects. What schema policy is typically needed? Insert fabricated values for every absent property without a rule Allow appropriate nullable fields while preserving missing versus malformed cases when material Flatten all objects into a single unparseable text blob for silver Reject every row missing any optional property Explanation Correct Answer B:Optional JSON attributes can be represented as nullable structured fields with explicit data quality handling. Incorrect Answers A:Invented defaults can create misleading analytics. C:Silver should preserve a meaningful structured representation when the source contract allows it. D:Optionality is part of the source contract and should not force universal rejection.
Question 608 A source CSV contains embedded double quotes represented by an escape convention. Ingest produces malformed descriptions. What reader property needs review? Only watermark for late events Only table auto optimize Only notebook scheduling timezone Quote escape behavior matching the actual CSV dialect Explanation Correct Answer D:Correct handling of embedded quotes depends on the source's escape and quoting convention. Incorrect Answers A:A late-event watermark does not control CSV escaping. B:Storage compaction does not change how source CSV quotes are parsed. C:Job triggers do not repair field tokenization.
Question 609 A data feed is copied from a source region to a landing region before Databricks ingestion. Compliance requires proving where each original row originated. What metadata should be kept? Only the Databricks cluster runtime version Only the landing bucket's name Only output row sorting Original source identity, region, delivery version, and immutable file or event reference Explanation Correct Answer D:Provenance fields should survive landing and transformation so compliance claims reflect the actual source, not just the landing bucket. Incorrect Answers A:Runtime version does not establish data origin. B:The landing location alone may conceal the original source system and region. C:Row order cannot provide source-region provenance.
Question 610 A partner supplies records as a Parquet dataset partitioned by `year` and `month`, and an analyst notices the partition columns do not appear as ordinary fields in each file. How should the reader recover them? Read the partitioned root using supported partition discovery and validate directory naming Assume partition columns must be manually inserted into every Parquet row Turn off schema reading entirely Delete every directory containing an equals sign Explanation Correct Answer A:Spark can infer partition columns from appropriate directory structure and path context, even when those values are not repeated inside each file. Incorrect Answers B:Well-formed partitioned datasets do not require duplicating partition keys in each data file. C:Parquet schema and partition discovery are needed to interpret the dataset reliably. D:Hive-style partition folders use key=value names as metadata rather than invalid files.