Messages by Thread
-
[I] Derive a native UDF's return type from the library instead of requiring the caller to declare it [datafusion-comet]
via GitHub
-
[PR] fix: aggregate dynamic filtering with unsupported expressions [datafusion]
via GitHub
-
[I] `round` on float/double falls back to Spark, while `bround` already uses the codegen dispatcher [datafusion-comet]
via GitHub
-
[PR] bench: `multi_group_by` cases to exercise `List` [datafusion]
via GitHub
-
[PR] docs: fix 'sqllite' typo in sqllogictest README [datafusion]
via GitHub
-
[I] `named_struct` with duplicate field names falls back to Spark [datafusion-comet]
via GitHub
-
[I] `lpad` / `rpad` with a non-literal `pad` argument falls back to Spark [datafusion-comet]
via GitHub
-
[PR] feat: route `round` on float/double through the codegen dispatcher [datafusion-comet]
via GitHub
-
[I] `translate` falls back to Spark by default instead of using the codegen dispatcher like the other string functions [datafusion-comet]
via GitHub
-
[I] NestedLoopJoin buffers the build side into a single concat_batches allocation: 2x transient peak, invisible to the memory pool [datafusion]
via GitHub
-
[I] Bug triage results: 2026-08-31 [datafusion-comet]
via GitHub
-
[PR] Non breaking cast metadata [datafusion]
via GitHub
-
[PR] feat: support unicode case sensitive field names for reading parquet [datafusion-comet]
via GitHub
-
[PR] chore: bench additional scalar functios [datafusion-comet]
via GitHub
-
[I] Align Comet's native UDF surface with SPARK-55278's language-agnostic UDF protocol [datafusion-comet]
via GitHub
-
[PR] Add ComposedNamedPhysicalExtensionCodec [datafusion]
via GitHub
-
Re: [PR] fix: record `join_time` and output metrics in PiecewiseMergeJoin [datafusion]
via GitHub
-
[I] SQLQueryTestSuite.ignoreList entries skip the Spark baseline, not just Comet [datafusion-comet]
via GitHub
-
Re: [I] Support Spark-compatible Unicode case-insensitive Parquet field matching [datafusion-comet]
via GitHub
-
[PR] feat(spark): add equal_null [datafusion]
via GitHub
-
[I] `length` / `bit_length` / `octet_length` fall back to Spark on binary input [datafusion-comet]
via GitHub
-
[I] Map lookups with float, collated or complex keys fall back to Spark (`map_col[key]`, `element_at`) [datafusion-comet]
via GitHub
-
[I] Route unrecognized `StaticInvoke` and `Invoke` through the codegen dispatcher instead of falling back [datafusion-comet]
via GitHub
-
Re: [I] Scalar NaN and signed-zero sort keys split WindowGroupLimit rank peers [datafusion-comet]
via GitHub
-
Re: [PR] Snowflake: Add EXTERNAL VOLUME DDL [datafusion-sqlparser-rs]
via GitHub
-
Re: [PR] feat: cleanup & simplify ballista config [datafusion-ballista]
via GitHub
-
Re: [PR] feat: custom Rust UDFs via arrow-ffi [experimental] [datafusion-comet]
via GitHub
-
[PR] chore: report the codegen-dispatch gate from getSupportLevel [datafusion-comet]
via GitHub
-
Re: [PR] Only apply non-left-associative nesting to deferred join constraints [datafusion-sqlparser-rs]
via GitHub
-
[PR] feat: drop provably-true IS NOT NULL conjuncts from FilterExec predicates via input null_count statistics [datafusion]
via GitHub
-
[PR] chore: move dead and defensive serde guards out of convert [datafusion-comet]
via GitHub
-
[I] Array functions fall back to Spark for binary and struct element types (`ArraysBase` type gate) [datafusion-comet]
via GitHub
-
Re: [I] Store job event logs in an object store so history survives the scheduler [datafusion-ballista]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
Re: [PR] feat: implement regr_slope, regr_intercept, regr_r2, regr_sxx, regr_syy, regr_sxy aggregates [datafusion-comet]
via GitHub
-
[I] Dictionary-encoded group keys pay 27-37% aggregation tax vs Utf8View in TPC-H (Q2/Q19/Q20), DF 54 & 55 both [datafusion]
via GitHub
-
[PR] Honour operator precedence in MySQL and Spark `DIV` [datafusion-sqlparser-rs]
via GitHub
-
[PR] chore: warn when a Compatible serde declines in convert [datafusion-comet]
via GitHub
-
[PR] chore: rename .claude directory to vendor-neutral .ai [datafusion-comet]
via GitHub
-
[I] Rename the .claude directory to a vendor-neutral .ai [datafusion-comet]
via GitHub
-
[I] Hash functions fall back to Spark for decimal precision > 18, and `sha2` for a non-literal `numBits` [datafusion-comet]
via GitHub
-
[I] `timestamp_seconds` falls back to Spark for decimal, byte and short input [datafusion-comet]
via GitHub
-
[I] `arrays_zip` falls back to Spark for map element types [datafusion-comet]
via GitHub
-
[I] `next_day` and `levenshtein` fall back to Spark on collated strings [datafusion-comet]
via GitHub
-
[I] `abs` on interval types falls back to Spark [datafusion-comet]
via GitHub
-
[I] `map_from_arrays` falls back to Spark under `mapKeyDedupPolicy=LAST_WIN`, unlike `map_from_entries` [datafusion-comet]
via GitHub
-
[I] `map_sort` falls back to Spark for non-scalar map key types [datafusion-comet]
via GitHub
-
[I] `to_csv` never runs inside Comet by default, unlike `to_json` / `from_csv` / `schema_of_csv` [datafusion-comet]
via GitHub
-
[I] Codegen dispatcher is unreachable from `convert`, so serdes that decline there never get a dispatch attempt [datafusion-comet]
via GitHub
-
[I] Codegen dispatcher: guard the closure-serialize step so a non-serializable tree falls back cleanly instead of throwing at plan time [datafusion-comet]
via GitHub
-
[PR] array_sort / array_has: monomorphize per native type instead of per Arrow type (reduce binary size) [datafusion]
via GitHub
-
[I] [EPIC] Codegen-dispatch coverage audit: expressions that fall back to Spark where the JVM dispatcher would work [datafusion-comet]
via GitHub
-
[I] Narrow the invalid-UTF-8 Comet opt-out in the sketch and hll SQL test files [datafusion-comet]
via GitHub
-
[I] Remove stale dev/diffs test exclusions whose tracking issues are closed [datafusion-comet]
via GitHub
-
Re: [I] `ceil`, `floor` on `decimal` can produce spurious overflow [datafusion]
via GitHub
-
Re: [I] RelationPlannerContext exposes no CTE scope, so extension relation planners silently shadow CTEs [datafusion]
via GitHub
-
Re: [PR] Support `CREATE VECTOR INDEX` [datafusion-sqlparser-rs]
via GitHub
-
[PR] perf: buffer the NestedLoopJoin build side as coalesced chunks instead of one concat_batches allocation [datafusion]
via GitHub
-
Re: [PR] Make extension codecs composable [datafusion-python]
via GitHub
-
Re: [PR] Fix final hash aggregate output regression by materializing once [datafusion]
via GitHub
-
Re: [PR] Avoid concatenating record batches in joins [datafusion]
via GitHub
-
[PR] Parse SQLite's `IS` and `IS NOT` operators [datafusion-sqlparser-rs]
via GitHub
-
[PR] Parse SQLite boolean PRAGMA values [datafusion-sqlparser-rs]
via GitHub
-
Re: [I] Increased memory usage due to round robin tie-breaker feature in SortPreservingMergeStream [datafusion]
via GitHub
-
[PR] Parse PostgreSQL MERGE DO NOTHING actions [datafusion-sqlparser-rs]
via GitHub
-
[PR] fix: disable aggregate dynamic filter when unsupported expressions are present (#24816) [datafusion]
via GitHub
-
[PR] perf: reuse per-partition scratch in the shuffle write path [datafusion-comet]
via GitHub
-
[I] Aggregate dynamic filter can prune rows required by unsupported expressions [datafusion]
via GitHub
-
Re: [I] [EPIC] Split Aggregation Logic into Dedicated Streams [datafusion]
via GitHub
-
[PR] perf: hold per-group accumulator state in blocks once it is large [datafusion]
via GitHub