87 real Delta Lake questions from the Big Data bank, as asked in Indian campus drives and tech interviews. Every question has a verified answer and an AI-tutor explanation on placd — free to start.
1. What is Deletion vectors?
Junior
A.a Delta table column whose values are computed from other columns by a stored expression, useful for deriving partition keys automatically
B.the Delta feature that runs a compaction pass after a write when many small files are detected, combining them automatically
C.a Delta Lake feature that marks deleted rows in a separate file so updates avoid rewriting entire Parquet files, a merge-on-read style
D.the tradeoff where UniForm exposes Delta tables to Iceberg readers without a second copy, versus maintaining a separate Iceberg table
2. Which term means: "a Delta Lake feature that marks deleted rows in a separate file so updates avoid rewriting entire Parquet files, a merge-on-read style"?
A.Deletion vectors — the Delta SQL command that removes rows matching a predicate transactionally and records the change in the table log
B.Deletion vectors — a Delta Lake feature that marks deleted rows in a separate file so updates avoid rewriting entire Parquet files, a merge-on-read style
C.Deletion vectors — reading a Delta table as a streaming input where each new committed version supplies the next micro-batch of records
D.Deletion vectors — the Delta feature that shuffles data before writing to produce fewer, appropriately sized files and reduce the small-file problem
A.Change Data Feed — reading a Delta table as a streaming input where each new committed version supplies the next micro-batch of records
B.Change Data Feed — the Delta behavior that rejects writes whose schema does not match the table unless schema evolution is explicitly allowed
C.Change Data Feed — a Delta Lake feature that records row-level inserts, updates, and deletes so downstream consumers can read only what changed
D.Change Data Feed — a Delta Lake data layout that automatically clusters data on chosen keys and adapts over time, replacing rigid partitioning and Z-ordering
8. Which term means: "a Delta Lake data layout that automatically clusters data on chosen keys and adapts over time, replacing rigid partitioning and Z-ordering"?
A.Liquid Clustering — a Delta Lake feature that generates Iceberg and Hudi metadata over Delta data so multiple engines can read one copy
B.Liquid Clustering — a Delta Lake data layout that automatically clusters data on chosen keys and adapts over time, replacing rigid partitioning and Z-ordering
C.Liquid Clustering — writing a stream into a Delta table with exactly-once guarantees enforced through a checkpoint and transaction log
D.Liquid Clustering — the Delta 4.0 capability to widen a column's type, such as int to long, without rewriting existing data files
A.Checkpoint file — a Delta clone that copies only metadata and references the source's existing data files, creating an inexpensive isolated table for testing
B.Checkpoint file — a Parquet summary of the Delta transaction log written periodically so readers avoid replaying every JSON commit
C.Checkpoint file — a Delta clone that copies both metadata and all data files, producing a fully independent table that survives source deletion
D.Checkpoint file — a Delta table column whose values are computed from other columns by a stored expression, useful for deriving partition keys automatically
A.Optimistic concurrency control — a Delta table column whose values are computed from other columns by a stored expression, useful for deriving partition keys automatically
B.Optimistic concurrency control — the Delta mechanism that lets concurrent writers proceed and resolves conflicts at commit by validating read and write sets
C.Optimistic concurrency control — the Delta feature that runs a compaction pass after a write when many small files are detected, combining them automatically
D.Optimistic concurrency control — a Delta clone that copies only metadata and references the source's existing data files, creating an inexpensive isolated table for testing
A.Schema enforcement — a Delta table column whose values are computed from other columns by a stored expression, useful for deriving partition keys automatically
B.Schema enforcement — a Delta clone that copies both metadata and all data files, producing a fully independent table that survives source deletion
C.Schema enforcement — the tradeoff where UniForm exposes Delta tables to Iceberg readers without a second copy, versus maintaining a separate Iceberg table
D.Schema enforcement — the Delta behavior that rejects writes whose schema does not match the table unless schema evolution is explicitly allowed
23. Which term means: "the tradeoff where UniForm exposes Delta tables to Iceberg readers without a second copy, versus maintaining a separate Iceberg table"?
A.Delta UniForm vs native Iceberg — the Delta SQL command that removes rows matching a predicate transactionally and records the change in the table log
B.Delta UniForm vs native Iceberg — the tradeoff where UniForm exposes Delta tables to Iceberg readers without a second copy, versus maintaining a separate Iceberg table
C.Delta UniForm vs native Iceberg — the Delta 4.0 capability to widen a column's type, such as int to long, without rewriting existing data files
D.Delta UniForm vs native Iceberg — the ordered _delta_log of JSON commits that gives a Delta table its ACID guarantees and time travel
A.Delta Lake 4.0 type widening — the Delta 4.0 capability to widen a column's type, such as int to long, without rewriting existing data files
B.Delta Lake 4.0 type widening — the tradeoff where UniForm exposes Delta tables to Iceberg readers without a second copy, versus maintaining a separate Iceberg table
C.Delta Lake 4.0 type widening — the Delta feature that shuffles data before writing to produce fewer, appropriately sized files and reduce the small-file problem
D.Delta Lake 4.0 type widening — the Delta merge-on-read feature marking removed rows in an auxiliary bitmap so deletes and updates avoid rewriting whole Parquet files
A.VACUUM safety window — the Delta merge-on-read feature marking removed rows in an auxiliary bitmap so deletes and updates avoid rewriting whole Parquet files
B.VACUUM safety window — a Parquet summary of the Delta transaction log written periodically so readers avoid replaying every JSON commit
C.VACUUM safety window — writing a stream into a Delta table with exactly-once guarantees enforced through a checkpoint and transaction log
D.VACUUM safety window — the retention interval that VACUUM honors to avoid deleting files still needed by in-flight readers or time travel
Showing 30 of 87 Delta Lake questions — the full set, with answers, explanations and an AI tutor on every question, is inside.
Free to start
Answers, AI explanations, and a scored voice mock interview
Sign up free to check your answers with explanations, ask the AI tutor anything on any question, and take one full AI mock interview — scored like a real panel.