Problemsby GlyphApache Iceberg users keep filing bugs
11 posts from 10 people9d activeSteady, +1 this week
Posts per day
The posts behind it
Two free hands-on badges from Snowflake Expedition (finance reconciliation + campaign planning), both built on Iceberg, CoCo and Cortex Agents
...Smarter: Build It with CoCo, Ask It with CoWork** (James Cha-Earley). You land three messy financial exports (payroll register, GL export, bank statement) in Snowflake-managed Apache Iceberg tables, use CoCo to clean, standardize and match transactions across systems, and show...
FileMetadata.Builder.copy(DeleteFile) drops equalityFieldIds
...then add it with `newRowDelta().addDeletes(...)` and run `newScan().planFiles()`. Workaround: call `.ofEqualityDeletes(ids)` after `.copy()`, as `RewriteTablePathUtil` already does.
`latest_sequence_number` is NULL for format v1 Iceberg snapshots instead of 0
...snapshots. - Joining `latest_sequence_number` to `system.iceberg_files.sequence_number` fails to match for any v1 table. Root cause: In `IcebergMetadata.cpp` the `set_latest_snapshot` lambda only fills `latest_sequence_number` when the snapshot JSON has a `sequence-number` key,...
Nessie: listNamespaces fails for namespace names containing ' or \
### Apache Iceberg version main (development) ### Query engine None ### Please describe the bug 🐞 **Problem** `NessieIcebergClient.listNamespaces` builds a CEL filter for child namespaces by putting the parent's name inside a single-quoted string literal, without escaping it:...
Temporal transforms put pre-epoch timestamps at `.999999` into the previous unit
### Apache Iceberg version main (development) ### Query engine Other ### Please describe the bug 🐞 `DateTimeUtil.convertMicros` and `DateTimeUtil.convertNanos` return the previous unit for a pre-epoch timestamp in the first second of a unit when its fraction is `.999999`...
Spec, OpenAPI: Define hex case for binary and fixed values
### Feature Request / Improvement The table spec and the OpenAPI spec disagree on how `binary` and `fixed` values are written as hex strings: - `format/spec.md`, Appendix D (JSON single-value serialization): "Stored as a hexadecimal string", with the example `"000102ff"`...
Nessie: listNamespaces checks the parent namespace at a stale ref hash
**Apache Iceberg version** main @ 9d7b2c52f **Query engine** None - the bug is in the Nessie catalog client, so it is engine-agnostic. **Please describe the bug** `NessieIcebergClient.listNamespaces(Namespace)` reads two different commits in one call, so listing a namespace...
MR: IcebergSplit.getLocations discards cached block locations on repeated calls
**Apache Iceberg version** main @ 9d7b2c52f **Query engine** None - the bug is in `iceberg-mr` and is engine-agnostic. **Please describe the bug** `IcebergSplit.getLocations()` (`mr/src/main/java/org/apache/iceberg/mr/mapreduce/IcebergSplit.java` line 70-82) returns the computed...
File Type Support
### Proposed Change Authors: Talat Uyarer, Alex Stephen, Sung Yun, Gaurav Soni Add file type to Iceberg v4 Add a new file primitive type to the Iceberg v4 spec for referencing a range of bytes, stored either inline or in an external object. The type and its value semantics are...
ORC: Writer accepts an empty struct that reader cannot bind
This issue was filed incorrectly and is withdrawn.
[RCK]
...to make the RCK use the remote REST endpoint instead of local storage. However, the two failing test cases still use local absolute paths. Is this expected behavior, or is there an additional configuration required to disable local path usage when running RCK against a remote...
Companies and products named
About this problem
Evidence
11 posts from 10 people in 3 places, about 1 a week over 68 days. Mostly on apache/iceberg, ClickHouse/ClickHouse, r/snowflake. Tools named alongside: GitHub, Amazon S3, Docker, Temporal.
Frustration Frustration 0 of 3· Seen on GitHub issues, Reddit
How it was grouped
Posts about Apache Iceberg that state a pain but match no known issue yet. First post Aug 31, 2026, latest Nov 6, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method
Other Apache Iceberg problems
Trouble with Apache Iceberg table
Top post: "Parquet files written with the default CRS are not compatible with Iceberg Java EPSG:4326 table", one of 12 posts from 8 people on GitHub issues.
Trouble with Apache Iceberg spark
Top post: "Spark: WAP snapshot isolation broken for format-version 3 tables when deletion vectors are used (spark.wap.id)", one of 7 posts from 7 people on GitHub issues.
Apache Iceberg upgrades break working setups
6 posts from 5 people on GitHub issues.
Apache Iceberg runs out of memory
4 posts from 3 people on GitHub issues.
Trouble with Apache Iceberg rest
Top post: "Core: Negative delete-file-references in REST scan response cause IndexOutOfBoundsException", one of 4 posts from 3 people on Reddit, GitHub issues.
Apache Iceberg schema changes break downstream pipelines
3 posts from 3 people on GitHub issues.
Apache Iceberg connectors and syncs are fragile
3 posts from 3 people on GitHub issues.
Trouble with Apache Iceberg delete
Top post: "Iceberg DROP PARTITION aborts or republishes parent delete counters after an overwrite snapshot", one of 3 posts from 3 people on GitHub issues.
Trouble with Apache Iceberg tables
Top post: "Iceberg tables are much slower than MergeTree in the performance tests: subcolumns, sort order, join", one of 3 posts from 3 people on GitHub issues.
Bad data slips through Apache Iceberg unnoticed
3 posts from 3 people on GitHub issues.
Apache Iceberg deprecations force unplanned migrations
2 posts from 2 people on GitHub issues. Unverified.
Trouble with Apache Iceberg pruning
Top post: "API: Predicate evaluation and file pruning treat -0.0 and 0.0 as different values", one of 2 posts from 2 people on GitHub issues. Unverified.
History