All issues

[DISCUSSION] Roadmap
#1001 · shehabgamin opened on Oct 30, 2025
7
[EPIC] Catalog
#980 · shehabgamin opened on Oct 21, 2025
[EPIC] Full Coverage of Spark DataFrame and Spark SQL Tests
#500 · shehabgamin opened on May 27, 2025

#2573

· jonasdedden opened

on Sep 12, 2026

MERGE INTO ... WHEN MATCHED THEN UPDATE duplicates the entire file containing a match, not just the matched rows

#2570

· KabirBhutti opened

on Sep 11, 2026

Iceberg partitioned writes fan every task into every partition, so open writers and file count scale with partition cardinality

#2563

· Robin-Everaars opened

on Sep 8, 2026

[draft] CoalesceExec over Delta MERGE writer input stalls non-zero output partitions

#2561

· zemin-piao opened

on Sep 8, 2026

Filtering a projected derived column fails with `No field named "?table?".resource`

#2560

· zemin-piao opened

on Sep 7, 2026

Background `SparkConnectServer` is shut down when the Python handle is garbage collected, causing a silent indefinite hang

#2530

· djouallah opened

on Sep 1, 2026

Cannot cast Timestamp(Millisecond/Nanosecond/Second, tz) to Spark type when analyzing Parquet schema

#2519

· amit-odaia opened

on Aug 27, 2026

`CASE`/`if` resolve to the type of the first branch instead of the least common type of all branches

#2503

· amit-odaia opened

on Aug 25, 2026

Python UDFs run inline on the primary runtime, so a saturating UDF batch starves the Spark Connect server

#2472

· james-willis opened

on Aug 21, 2026

SIGABRT on Linux when the idle-session reaper deletes a session while a Python UDF is in flight

#2471

· james-willis opened

on Aug 21, 2026

Python UDF execution destroys the PyThreadState after every call, breaking libraries with threading.local-tied native state (use-after-free in pyproj)

#2456

· james-willis opened

on Aug 19, 2026

DATE + sub-day INTERVAL returns DATE (drops time), Spark returns TIMESTAMP

#2370

· vwodaia opened

on Aug 5, 2026

Navigation Menu

Uh oh!

All issues

Lance format compatibility

MERGE INTO ... WHEN MATCHED THEN UPDATE duplicates the entire file containing a match, not just the matched rows

Iceberg partitioned writes fan every task into every partition, so open writers and file count scale with partition cardinality

[draft] CoalesceExec over Delta MERGE writer input stalls non-zero output partitions

Filtering a projected derived column fails with `No field named "?table?".resource`

Background `SparkConnectServer` is shut down when the Python handle is garbage collected, causing a silent indefinite hang

Cannot cast Timestamp(Millisecond/Nanosecond/Second, tz) to Spark type when analyzing Parquet schema

`CASE`/`if` resolve to the type of the first branch instead of the least common type of all branches

Python UDFs run inline on the primary runtime, so a saturating UDF batch starves the Spark Connect server

SIGABRT on Linux when the idle-session reaper deletes a session while a Python UDF is in flight

Python UDF execution destroys the PyThreadState after every call, breaking libraries with threading.local-tied native state (use-after-free in pyproj)

DATE + sub-day INTERVAL returns DATE (drops time), Spark returns TIMESTAMP

Uh oh!

Issues

Search results