Search lesson summaries, concept aliases and glossary definitions. Each lesson links to useful
prerequisites and its companion evidence. Follow foundations for a first pass, then return for
deeper mechanisms. Try the practice cases →
114 chapters and 27 glossary entries match.
- parquet · foundation
Locate the magic bytes, column data and footer before following a byte offset.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Map physical leaf types to logical values and optional or repeated fields.
Useful first: File Structure
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Follow horizontal record groups into independently stored leaf columns.
Useful first: File Structure
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Find the pages belonging to one leaf column in one row group.
Useful first: Row Groups
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Separate page headers, encoded bodies and optional CRC integrity checks.
Useful first: Column Chunks
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · deep
Compare plain values, dictionary indices, runs and delta representations before compression.
Useful first: Pages
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · deep
Distinguish encoded bytes, compressed page bodies and decoded memory.
Useful first: Encodings
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Reconstruct null lists, empty lists and repeated children from definition and repetition levels.
Useful first: Schema & Types
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Use trusted min/max bounds to exclude candidates without asserting that a value exists.
Useful first: Row Groups
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · deep
Probe split-block filters: absence excludes work; a possible match needs residual evaluation.
Useful first: Statistics
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · deep
Combine page bounds and row locations to plan selective reads and conservative fallbacks.
Useful first: Pages·Statistics
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · deep
Follow schema, row-group metadata and offsets from the file trailer.
Useful first: File Structure·Column Chunks
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- parquet · foundation
Open a local file and connect metadata, decoded records and application read traces.
Useful first: Footer
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Real bytes, encodings, nested values, statistics and local file inspection.
The reader supports documented encodings and codecs. Encryption and external column chunks are rejected.
- arrow · foundation
Place Arrow in the path from stored columns to computed answers.
Start here; no earlier lesson required.
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Follow a logical row into typed values, validity, offsets and children.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Distinguish the physical slot from its logical null value using validity bits.
Useful first: A record batch
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Address original integer slots without losing precision.
Useful first: A record batch
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Use byte offsets for strings and child-slot offsets for lists, including empty and null lists.
Useful first: Validity
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Follow parent offsets into child arrays and their validity.
Useful first: A record batch
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · deep
Resolve dictionary indices against a separate values array.
Useful first: A record batch
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Apply a logical slice offset without confusing parent positions and child offsets.
Useful first: Variable-width strings
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · deep
Keep exact 64-bit integers when a display or JavaScript number would lose precision.
Useful first: A record batch
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Compare IPC file and stream framing around schemas, dictionaries and record batches.
Useful first: A record batch·Dictionary arrays
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Compare the storage and memory representations of the same values.
Useful first: A record batch
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Inspect original IPC files and reproduce the pinned native buffer observations.
Useful first: IPC files and streams
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- arrow · foundation
Apply the Arrow walkthrough to an original artefact in the local tools.
Useful first: What Arrow is
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Native typed arrays, validity and offsets, shared-buffer slices, IPC file/stream batches and a bounded local inspector.
The guide decodes a pinned fixture. The IPC tool uses Apache Arrow JavaScript 17.0.0; it is not a table reader or a constant-memory scanner.
- iceberg · foundation
Place Iceberg in the path from stored columns to computed answers.
Start here; no earlier lesson required.
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Select table rows through a committed snapshot rather than every file in a directory.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Trace metadata JSON to a manifest list, manifests and original Parquet data.
Useful first: The table is a selection.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Read file entries with partition values, statistics and data or delete content.
Useful first: Every arrow has evidence.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Interpret existing, added and deleted entries in their snapshot context.
Useful first: A manifest is a file inventory with meaning.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Follow a new commit that adds files while preserving earlier file identity.
Useful first: The table is a selection.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Compare committed memberships and answers across snapshots.
Useful first: New files join. Old files keep their identity.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Follow a copy-on-write deletion that replaces a data file.
Useful first: Change the snapshot, change the table.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Separate equality keys from file-path and physical-position deletes.
Useful first: Remove one row. Rewrite its file.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Inspect original delete files and connect their values to affected data rows.
Useful first: Column Chunks·Two kinds of delete file, two different identities.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Reason about delete applicability before evaluating a remaining-row predicate.
Useful first: Two kinds of delete file, two different identities.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Test sequence, field identity, partition specification and rewrite boundaries for deletes.
Useful first: Two kinds of delete file, two different identities.·Partition evolution changes new writes, not old files.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Follow cumulative v3 deletion vectors and the positions retained from earlier deletes.
Useful first: Two kinds of delete file, two different identities.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Project predicates into partition and file bounds without inventing matches.
Useful first: Statistics·A manifest is a file inventory with meaning.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Read old and new partition layouts using each file’s own specification.
Useful first: Rule out files before opening them.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Track stable field IDs through a rename while preserving the meaning of historical data.
Useful first: The table is a selection.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Distinguish a renamed field from a new field that reuses an old name.
Useful first: A field’s name can change. Its identity stays.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Separate staged files from the atomic metadata-pointer update that publishes a snapshot.
Useful first: New files join. Old files keep their identity.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Trace retained references before deciding whether an old file is unreachable.
Useful first: Change the snapshot, change the table.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Compare native append and conflicting rewrite commits against a stale snapshot.
Useful first: The pointer update is the commit.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · deep
Distinguish snapshot expiration from orphan cleanup using recorded native outcomes.
Useful first: Removing a snapshot reference does not make every old file garbage.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Open an original physical Parquet file while preserving its table-membership context.
Useful first: Row Groups·The table is a selection.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · reference
Download original metadata and data, pinned recipes and independent observations.
Useful first: Every arrow has evidence.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- iceberg · foundation
Apply the Iceberg walkthrough to an original artefact in the local tools.
Useful first: What Iceberg is
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Snapshots, field identity, native equality/position deletes, conflicting writers, cleanup and cumulative v3 deletion vectors.
Curated tables and native experiments. The browser is not a general Iceberg reader or a remote catalog client.
- delta · foundation
Place Delta Lake in the path from stored columns to computed answers.
Start here; no earlier lesson required.
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Explain why the files on disk contain more rows than the active table.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Apply ordered transaction-log actions to reconstruct active membership.
Useful first: Three files. Twenty-three rows. A seventeen-row table.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Follow a replacement file and remove action for a copy-on-write delete.
Useful first: Read the log. Watch the table take shape.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Compare logical answers at three committed versions.
Useful first: Read the log. Watch the table take shape.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Interpret add and remove actions without treating removed data as immediately erased.
Useful first: Read the log. Watch the table take shape.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Separate active data files from historical files still present on disk.
Useful first: Actions change membership.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Start replay from materialized state and apply subsequent log actions.
Useful first: Read the log. Watch the table take shape.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Use active-file statistics to exclude work while retaining uncertain candidates.
Useful first: Statistics·Actions change membership.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · deep
Inspect native deletion-vector bytes and cumulative masks over original physical rows.
Useful first: Delete one row. Replace one file.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · deep
Follow native conflicts, retry validation and recomputation before publication.
Useful first: Actions change membership.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · deep
Observe native mask materialization and cleanup of superseded physical files.
Useful first: Original deletion vectors and cumulative masks
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · deep
Model the difference between staging files and publishing a valid optimistic commit.
Useful first: Actions change membership.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · deep
Separate reconstructable historical membership from the continued availability of its data files.
Useful first: Three commits, three answers.·A checkpoint changes where replay begins.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · deep
Check reader feature requirements before claiming support for a Delta table.
Useful first: Read the log. Watch the table take shape.·Original deletion vectors and cumulative masks
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Compare log replay and manifest-based membership across table formats.
Useful first: The table is a selection.·Read the log. Watch the table take shape.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · reference
Reproduce the original log, checkpoint, data files and native lifecycle observations.
Useful first: A checkpoint changes where replay begins.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- delta · foundation
Apply the Delta Lake walkthrough to an original artefact in the local tools.
Useful first: What Delta Lake is
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Legacy log/checkpoint replay plus native deletion vectors, writer conflicts, delete materialization and physical retention cleanup.
The general lesson replayer accepts reader 1 / writer 2. Advanced native examples do not imply arbitrary Delta feature support.
- duckdb · foundation
Place DuckDB in the path from stored columns to computed answers.
Start here; no earlier lesson required.
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Follow a SQL question through scan, filter, projection and aggregation.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Map a predicate’s selected positions back to an input vector.
Useful first: One question, a few operators.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · deep
Compare flat, constant and dictionary vector representations.
Useful first: Eighteen values enter. Which five leave?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · deep
Compose a second selection through the first selection’s row mapping.
Useful first: Eighteen values enter. Which five leave?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Inspect result batches and their Arrow memory buffers.
Useful first: A record batch·Eighteen values enter. Which five leave?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Execute a bundled query in the browser and distinguish it from recorded native results.
Useful first: One question, a few operators.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Read the recorded physical operator tree and its measured cardinalities.
Useful first: One question, a few operators.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Attach values through key equality while preserving duplicate matches.
Useful first: Eighteen values enter. Which five leave?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Maintain sum and count state across batches to compute a final average.
Useful first: Eighteen values enter. Which five leave?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · deep
Explain why LIMIT caps output while a filter, sort or aggregate may consume more input.
Useful first: An average is a sum and a count in progress.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · deep
Compare pruning for identical values written in different physical orders.
Useful first: Statistics·The actual operator tree.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · reference
Reproduce the pinned native query results, plans and buffer observations.
Useful first: The actual operator tree.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- duckdb · foundation
Apply the DuckDB walkthrough to an original artefact in the local tools.
Useful first: What DuckDB is
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Live local SELECT/CTE and EXPLAIN; native plans, typed result buffers, joins, partial aggregation and physical locality.
The guide uses recorded examples. The SQL tool runs separately with explicit input, time and result limits.
- engines · foundation
Place a query engine in the path from stored columns to computed answers.
Start here; no earlier lesson required.
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Follow a join query from input rows to an ordered result.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Move a legal filter before a join to reduce intermediate work.
Useful first: Follow the join.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Build a hash lookup and probe candidate keys, preserving duplicate multiplicity.
Useful first: Follow the join.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Resolve hash collisions with equality checks, including duplicate and null keys.
Useful first: Build and probe
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Predict when moving a filter across an outer join changes null-extended rows.
Useful first: Why attach a description to a row we will discard?·Build and probe
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Distinguish streaming output from blocking state that needs more input.
Useful first: Follow the join.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Explain why keeping the smallest three values can still require scanning the last row.
Useful first: A pipeline is constrained by what each operator must know.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Merge sorted local runs to establish global order.
Useful first: Keeping the smallest three still requires seeing the last row.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Connect a final ordered result to its typed Arrow buffers.
Useful first: A record batch·Local order is not global order.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Locate a null’s physical slot and validity bit in an output batch.
Useful first: Validity
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Compare recorded DuckDB and DataFusion operator plans for the same query.
Useful first: Follow the join.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Reason about moving projections and filters while preserving semantics.
Useful first: Why attach a description to a row we will discard?·Compare the actual plans.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Inspect native memory-limit failures, spill evidence and LIMIT input work.
Useful first: A pipeline is constrained by what each operator must know.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · deep
Challenge cardinality estimates, casts and join order with recorded counterexamples.
Useful first: Optimization moves work toward its input.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · reference
Download pinned native plans, results and reproduction scripts.
Useful first: Compare the actual plans.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- engines · foundation
Apply the query-engine comparison to an original artefact in the local tools.
Useful first: What a query engine is
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
DuckDB/DataFusion plans and exact results, build/probe, join legality, ordering, skew, Arrow buffers and native memory/spill profiles.
Plan captures and explanatory playback do not measure current browser scheduling, device I/O or comparative performance.
- lance · foundation
Place Lance in the path from stored columns to computed answers.
Start here; no earlier lesson required.
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Compute exact neighbors under a specified distance rule.
Start here; no earlier lesson required.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Account for each candidate’s distance before ranking neighbors.
Useful first: Which vectors are closest?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · deep
Compare exact search with a native IVF_FLAT index and measure missed neighbors as probes change.
Useful first: Make every candidate account for its distance.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Group selected row requests to see when physical reads can be shared.
Useful first: Which vectors are closest?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Follow a dataset version’s fragment and data-file references.
Useful first: Which vectors are closest?
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Decode footer locations and follow the next range into column metadata.
Useful first: The manifest connects the table.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Connect known file regions to original bytes in the unindexed format-2.1 fixture.
Useful first: The manifest connects the table.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Translate a selected row address into the fixture’s physical byte range.
Useful first: Follow one file's byte ranges.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Separate the IDs chosen by search from storage work needed to fetch their values.
Useful first: Three nearby vectors can share the work of one.·From a row to its bytes
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · deep
Track logical IDs, physical addresses and index coverage through append, delete and compaction.
Useful first: An index chooses which distances will never be evaluated.·From a row to its bytes
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · deep
Follow fixed-width and variable-width values across different page boundaries.
Useful first: The footer tells a reader where to ask the next question.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Distinguish nearest-neighbor selection, dataset membership and physical file layout.
Useful first: Search selects IDs; storage supplies values.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · reference
Download the original unindexed file separately from native index experiments and recipes.
Useful first: Keep the layers distinct.
Evidence and topic tools
Companion originals, recorded versions and reproduction recipes:
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.
- lance · foundation
Apply the Lance walkthrough to an original artefact in the local tools.
Useful first: What Lance is
Evidence and topic tools
This explanation uses the guide’s illustrated examples. It has no separate native evidence
family.
Original format-2.1 fragments and buffers, exact and IVF_FLAT search, native index lifecycle, and multi-page nullable/compressed layouts with bounded ID decoding.
The examples are separate datasets. Nullable strings are native SDK results; bounded decoders and recorded queries do not constitute a general Lance reader or live ANN service.