Iceberg interview questions

1. A.an Iceberg spec v3 feature assigning each row a stable _row_id and sequence number to support change tracking
2. B.an Iceberg feature where partition values are derived from columns via transforms, so queries need not filter on a separate partition column
3. C.an Iceberg merge-on-read delete file marking removed rows by their file path and row position, efficient to apply but tied to specific data files
4. D.a named reference to a specific Iceberg snapshot, often used to mark a release or retain a snapshot for a retention period

Answer + AI explanation with Pro

1. A.rewrite manifests
2. B.sort order
3. C.Partition evolution
4. D.Hidden partitioning

Answer + AI explanation with Pro

1. A.Hidden partitioning — an Iceberg virtual table (such as files, snapshots, history, or partitions) queried to inspect a table's structure and evolution
2. B.Hidden partitioning — an Iceberg feature where partition values are derived from columns via transforms, so queries need not filter on a separate partition column
3. C.Hidden partitioning — the Iceberg catalog that adds Git-like branches, tags, and commits over tables so multi-table changes can be versioned and merged
4. D.Hidden partitioning — an Iceberg merge-on-read delete file marking removed rows by their file path and row position, efficient to apply but tied to specific data files

Answer + AI explanation with Pro

1. A.the Iceberg ability to change a table's partition scheme over time without rewriting existing data files
2. B.an Iceberg merge-on-read delete file marking removed rows by their file path and row position, efficient to apply but tied to specific data files
3. C.the Iceberg SQL statement that performs upserts by matching a source against a target and inserting, updating, or deleting in one atomic operation
4. D.the open-source REST Iceberg catalog originally open sourced by Snowflake, adding role-based access control and multi-engine governance

Answer + AI explanation with Pro

1. A.Partition evolution
2. B.Merge-on-read
3. C.compaction
4. D.REST catalog

Answer + AI explanation with Pro

1. A.Partition evolution — the Iceberg row-level mode that records changes as delete and data files merged at query time, favoring fast writes over fast reads
2. B.Partition evolution — the Iceberg ability to change a table's partition scheme over time without rewriting existing data files
3. C.Partition evolution — an Iceberg update mode that rewrites whole data files on each change, giving fast reads but slower writes
4. D.Partition evolution — an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files

Answer + AI explanation with Pro

1. A.an Iceberg update mode that rewrites whole data files on each change, giving fast reads but slower writes
2. B.an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files
3. C.an immutable, point-in-time version of an Iceberg table listing all data files, enabling time travel and rollback
4. D.an Iceberg table property declaring how rows should be ordered within data files so writers cluster data and readers prune more effectively

Answer + AI explanation with Pro

1. A.write distribution mode
2. B.sort order
3. C.Snapshot
4. D.REST catalog

Answer + AI explanation with Pro

1. A.Snapshot — the Iceberg SQL statement that performs upserts by matching a source against a target and inserting, updating, or deleting in one atomic operation
2. B.Snapshot — an immutable, point-in-time version of an Iceberg table listing all data files, enabling time travel and rollback
3. C.Snapshot — the Iceberg property (none, hash, or range) controlling how Spark redistributes rows across tasks before writing to limit small files and skew
4. D.Snapshot — an Iceberg update mode that records deletes and updates separately and merges them at query time for faster writes

Answer + AI explanation with Pro

1. A.the Iceberg row-level mode that records changes as delete and data files merged at query time, favoring fast writes over fast reads
2. B.an Iceberg metadata file that lists data files with their partition values and column statistics for pruning
3. C.an Iceberg update mode that records deletes and updates separately and merges them at query time for faster writes
4. D.an Iceberg merge-on-read delete file marking removed rows by their file path and row position, efficient to apply but tied to specific data files

Answer + AI explanation with Pro

1. A.merge-on-read
2. B.Apache Polaris
3. C.copy-on-write
4. D.Manifest file

Answer + AI explanation with Pro

1. A.Manifest file — an Iceberg table property declaring how rows should be ordered within data files so writers cluster data and readers prune more effectively
2. B.Manifest file — an Iceberg merge-on-read delete file marking removed rows by their file path and row position, efficient to apply but tied to specific data files
3. C.Manifest file — an Iceberg metadata file that lists data files with their partition values and column statistics for pruning
4. D.Manifest file — an Iceberg update mode that records deletes and updates separately and merges them at query time for faster writes

Answer + AI explanation with Pro

1. A.an open-source Iceberg REST catalog (1.0 in 2025) providing centralized governance and credential vending across engines
2. B.a named, independent line of Iceberg snapshots that supports isolated writes such as a write-audit-publish staging area
3. C.an Iceberg metadata file that lists data files with their partition values and column statistics for pruning
4. D.the Iceberg property (none, hash, or range) controlling how Spark redistributes rows across tasks before writing to limit small files and skew

Answer + AI explanation with Pro

1. A.Polaris catalog
2. B.Branch
3. C.write distribution mode
4. D.merge-on-read

Answer + AI explanation with Pro

1. A.Branch — an Iceberg merge-on-read delete file marking removed rows by column values such as id equals five, flexible to write but costly to apply on read
2. B.Branch — an Iceberg feature where partition values are derived from columns via transforms, so queries need not filter on a separate partition column
3. C.Branch — a named, independent line of Iceberg snapshots that supports isolated writes such as a write-audit-publish staging area
4. D.Branch — an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files

Answer + AI explanation with Pro

1. A.an Iceberg catalog interface defined by a standard HTTP API so engines can share one catalog implementation across vendors
2. B.the Iceberg rewrite_data_files action that merges many small files into larger ones and can re-sort data to restore read performance
3. C.an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files
4. D.a named reference to a specific Iceberg snapshot, often used to mark a release or retain a snapshot for a retention period

Answer + AI explanation with Pro

1. A.rewrite manifests
2. B.Snapshot
3. C.Tag
4. D.metadata table

Answer + AI explanation with Pro

1. A.Tag — the Iceberg SQL statement that performs upserts by matching a source against a target and inserting, updating, or deleting in one atomic operation
2. B.Tag — an Iceberg feature where partition values are derived from columns via transforms, so queries need not filter on a separate partition column
3. C.Tag — a named reference to a specific Iceberg snapshot, often used to mark a release or retain a snapshot for a retention period
4. D.Tag — the Iceberg catalog that adds Git-like branches, tags, and commits over tables so multi-table changes can be versioned and merged

Answer + AI explanation with Pro

1. A.an Iceberg update mode that rewrites whole data files on each change, giving fast reads but slower writes
2. B.the Iceberg ability to query a table as of an earlier snapshot or timestamp for reproducibility and auditing
3. C.the Iceberg catalog implementation that tracks table pointers in a Hive Metastore, a common bridge for existing Hadoop deployments
4. D.the Iceberg ability to change a table's partition scheme over time without rewriting existing data files

Answer + AI explanation with Pro

1. A.Copy-on-write
2. B.write distribution mode
3. C.Apache Polaris
4. D.Nessie catalog

Answer + AI explanation with Pro

1. A.Copy-on-write — an Iceberg update mode that rewrites whole data files on each change, giving fast reads but slower writes
2. B.Copy-on-write — an immutable, point-in-time version of an Iceberg table listing all data files, enabling time travel and rollback
3. C.Copy-on-write — an Iceberg virtual table (such as files, snapshots, history, or partitions) queried to inspect a table's structure and evolution
4. D.Copy-on-write — an Iceberg merge-on-read delete file marking removed rows by column values such as id equals five, flexible to write but costly to apply on read

Answer + AI explanation with Pro

1. A.a named, independent line of Iceberg snapshots that supports isolated writes such as a write-audit-publish staging area
2. B.an Iceberg update mode that records deletes and updates separately and merges them at query time for faster writes
3. C.an Iceberg update mode that rewrites whole data files on each change, giving fast reads but slower writes
4. D.an Iceberg merge-on-read delete file marking removed rows by their file path and row position, efficient to apply but tied to specific data files

Answer + AI explanation with Pro

1. A.Merge-on-read
2. B.Deletion vector
3. C.Time travel
4. D.Manifest file

Answer + AI explanation with Pro

1. A.Merge-on-read — an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files
2. B.Merge-on-read — an immutable, point-in-time version of an Iceberg table listing all data files, enabling time travel and rollback
3. C.Merge-on-read — an Iceberg metadata file that lists data files with their partition values and column statistics for pruning
4. D.Merge-on-read — an Iceberg update mode that records deletes and updates separately and merges them at query time for faster writes

Answer + AI explanation with Pro

1. A.an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files
2. B.an Iceberg update mode that rewrites whole data files on each change, giving fast reads but slower writes
3. C.an Iceberg spec v3 feature assigning each row a stable _row_id and sequence number to support change tracking
4. D.the Iceberg ability to query a table as of an earlier snapshot or timestamp for reproducibility and auditing

Answer + AI explanation with Pro

1. A.Deletion vector
2. B.Hive catalog
3. C.MERGE INTO
4. D.rewrite manifests

Answer + AI explanation with Pro

1. A.Deletion vector — the Iceberg rewrite_data_files action that merges many small files into larger ones and can re-sort data to restore read performance
2. B.Deletion vector — the Iceberg catalog implementation that stores table metadata pointers in the AWS Glue Data Catalog, the most adopted Iceberg catalog as of 2025
3. C.Deletion vector — the open-source REST Iceberg catalog originally open sourced by Snowflake, adding role-based access control and multi-engine governance
4. D.Deletion vector — an Iceberg spec v3 binary structure that marks deleted rows compactly per data file, replacing positional delete files

Answer + AI explanation with Pro

1. A.the Iceberg row-level mode that records changes as delete and data files merged at query time, favoring fast writes over fast reads
2. B.an Iceberg merge-on-read delete file marking removed rows by column values such as id equals five, flexible to write but costly to apply on read
3. C.the Iceberg SQL statement that performs upserts by matching a source against a target and inserting, updating, or deleting in one atomic operation
4. D.an Iceberg spec v3 feature assigning each row a stable _row_id and sequence number to support change tracking

Answer + AI explanation with Pro

1. A.Row lineage
2. B.Partition evolution
3. C.Hidden partitioning
4. D.metadata table

Answer + AI explanation with Pro

1. A.Row lineage — the open-source REST Iceberg catalog originally open sourced by Snowflake, adding role-based access control and multi-engine governance
2. B.Row lineage — an Iceberg spec v3 feature assigning each row a stable _row_id and sequence number to support change tracking
3. C.Row lineage — the Iceberg catalog that adds Git-like branches, tags, and commits over tables so multi-table changes can be versioned and merged
4. D.Row lineage — an Iceberg merge-on-read delete file marking removed rows by column values such as id equals five, flexible to write but costly to apply on read

Answer + AI explanation with Pro