Messages by Thread
-
[PR] feat(python): Add Python UDF support to Ballista Python bindings [datafusion-ballista]
via GitHub
-
Re: [PR] fix: account for row duplication and padding in join null counts [datafusion]
via GitHub
-
[PR] fix: return NULL from bit_xor when a sliding frame has no non-NULL values [datafusion]
via GitHub
-
[PR] bench: add ASOF join benchmark with a large unfiltered right input [datafusion]
via GitHub
-
[PR] SQLite: Accept fallback keywords as implicit column aliases [datafusion-sqlparser-rs]
via GitHub
-
[I] EPIC: Separate optimizer rules from correctness transformation rules [datafusion]
via GitHub
-
Re: [I] Invalid parquet writer options (e.g. `max_row_group_size = 0`, `bloom_filter_fpp = 1.5`) panic in `COPY` instead of returning an error [datafusion]
via GitHub
-
[PR] fix: forward `create_sliding_accumulator` to the inner UDAF when aliased [datafusion]
via GitHub
-
[PR] docs: Explain the role of `EnsureRequirements` in the physical optimizer pipeline [datafusion]
via GitHub
-
Re: [I] Add a GroupsAccumulator for the native regr_* aggregates [datafusion-comet]
via GitHub
-
Re: [PR] Improve UI [datafusion-site]
via GitHub
-
[I] Support hash-partitioned execution in AsOfJoinExec for memory-bounded joins [datafusion]
via GitHub
-
Re: [I] Optimize the evaluation of `DATE_TRUNC(<col>) == <constant>)` when pushed down [datafusion]
via GitHub
-
Re: [PR] Fix local Pelican build [datafusion-site]
via GitHub
-
[PR] fix: plan ORDER BY on a source column hidden by a select alias [datafusion]
via GitHub
-
Re: [PR] perf: preserve dictionary encoding through regexp_replace [datafusion]
via GitHub
-
[PR] perf: don't build a parquet stream for a file range with no row group left [datafusion]
via GitHub
-
[PR] SQLite: Support `IN` with a table name or table-valued function [datafusion-sqlparser-rs]
via GitHub
-
[PR] Refactoring: Merge the `IN` expression variants into `Expr::In` [datafusion-sqlparser-rs]
via GitHub
-
Re: [PR] ci: run the Rust workflow for fork pull requests on the fork [datafusion]
via GitHub
-
[I] Python client: plug into DataFusion Python natively through a custom FFI query planner [datafusion-ballista]
via GitHub
-
[I] Parquet scans still set up a reader for files whose row groups are all pruned [datafusion]
via GitHub
-
[PR] docs: fix missing word in BytesLikeIndexMap field comment [datafusion]
via GitHub
-
[PR] perf: deduplicate found expressions with an IndexSet [datafusion]
via GitHub
-
[I] Planning a query with many aggregate or window expressions slows down quadratically [datafusion]
via GitHub
-
[I] `corr`, `covar_pop`, `covar_samp` and `regr_*` silently ignore `DISTINCT` [datafusion]
via GitHub
-
[I] `bit_xor` over a sliding window frame returns 0 instead of NULL when the frame has only NULLs [datafusion]
via GitHub
-
[PR] SQLite: Bind <, <=, > and >= tighter than = [datafusion-sqlparser-rs]
via GitHub
-
Re: [I] alter_fp_rounding_mode Is Not Panic Safe [datafusion]
via GitHub
-
[I] Internal error: CAST to an extension-type field over a non-null column, under an aggregate (logical vs physical nullability) [datafusion]
via GitHub
-
Re: [PR] feat: rewrite count(non-nullable column) to count(1) [datafusion]
via GitHub
-
Re: [I] Wrong results: window aggregates over `RANGE ... N PRECEDING` depend on batch boundaries when the ORDER BY value is NULL [datafusion]
via GitHub
-
[I] docs: update Scala comments that reference the removed iceberg-writes.md page [datafusion-comet]
via GitHub
-
[PR] feat: support `[NOT] [I]LIKE ANY (...)` with a list of patterns [datafusion]
via GitHub
-
[PR] feat: support `RLIKE` and `REGEXP` operators [datafusion]
via GitHub
-
[PR] SQLite, DuckDB, PostgreSQL: Accept `._name` after whitespace in a qualified name [datafusion-sqlparser-rs]
via GitHub
-
[I] Aliased aggregate functions fail in sliding window frames [datafusion]
via GitHub
-
[PR] fix(functions): pass correct number_rows in gcd benchmarks [datafusion]
via GitHub
-
Re: [PR] docs: add NestedLoopJoinExec metrics to user-guide metrics page [datafusion]
via GitHub
-
[PR] fix: coerce INT96 in two-level and legacy Parquet lists [datafusion]
via GitHub
-
[I] coerce_int96 leaves INT96 in two-level and legacy Parquet lists as nanoseconds [datafusion]
via GitHub
-
[I] Draft: duplicate output names (decision 3) [datafusion]
via GitHub
-
Re: [I] Support columns having the same alias [datafusion]
via GitHub
-
Re: [PR] feat: execute scalar Arrow Python UDFs in native Comet [datafusion-comet]
via GitHub
-
Re: [PR] perf(sort-merge): cache current-row key per cursor to speed up SortPreservingMerge [datafusion]
via GitHub
-
[I] Physical plan check does not cover appended optimizer rules and does not name the rule that broke the plan [datafusion]
via GitHub
-
[PR] Check the final physical plan in the planner and name the rule that broke it [datafusion]
via GitHub
-
[PR] fix: keep a sort in EnsureRequirements when the ordering cannot be proven [datafusion]
via GitHub
-
[I] Scans of tables with a declared ordering use count-based file groups, so uneven files make them up to 3x slower [datafusion]
via GitHub
-
[PR] perf: balance file groups by size for scans with an output ordering [datafusion]
via GitHub
-
[PR] feat: write top-level Iceberg unknown columns natively [datafusion-comet]
via GitHub
-
Re: [PR] test(scheduler): freeze the plan each TPC-H query ends on, under both planners [datafusion-ballista]
via GitHub
-
Re: [PR] perf(scheduler): run one stage when the plan asks for the same stage twice [datafusion-ballista]
via GitHub
-
Re: [PR] parquet: cache per-file ArrowReaderMetadata for wide-schema scans [datafusion]
via GitHub
-
Re: [PR] feat: extend interval arithmetic to support scalar functions (UDFs) [datafusion]
via GitHub
-
Re: [PR] Split the partitioned hash join dynamic filter into bounds AND membership [datafusion]
via GitHub
-
Re: [PR] fix: scope HashJoinExec build_time to hash-table construction only [datafusion]
via GitHub
-
Re: [I] Decorrelate subqueries whose grouping sets leave out the correlated column [datafusion]
via GitHub
-
Re: [I] AQE: resolve_stage and rollback_*_stage drop the stage from the graph when the conversion fails [datafusion-ballista]
via GitHub
-
Re: [I] Prune files by statistics in `ListingTable` before building file groups [datafusion]
via GitHub
-
[PR] fix: handle NullType struct fields in JVM columnar shuffle [datafusion-comet]
via GitHub
-
Re: [PR] feat: release the output of a materialized CTE when the last scan finishes [datafusion]
via GitHub
-
[PR] feat: stream the output of a materialized CTE to its scans [datafusion]
via GitHub
-
[PR] PostgreSQL, ClickHouse: Keep custom data type modifiers re-parseable with a typed list [datafusion-sqlparser-rs]
via GitHub
-
[PR] fix: keep per-partition constant sort keys in a partitioned TopK [datafusion]
via GitHub
-
[PR] fix: point the spill limit error at spark.comet.maxTempDirectorySize [datafusion-comet]
via GitHub
-
[I] Wrong results: UNION ALL ... ORDER BY ... LIMIT when a sort key is constant per arm (partitioned TopK shares a threshold over a shortened key) [datafusion]
via GitHub
-
[PR] PostgreSQL: Parse an empty SELECT target list without FROM [datafusion-sqlparser-rs]
via GitHub
-
[PR] replaces the hard-coded dialect list in the tokenizer with a `supports_hash_comment` dialect method [datafusion-sqlparser-rs]
via GitHub
-
[I] ClickHouse: # comments are not recognized [datafusion-sqlparser-rs]
via GitHub
-
[PR] PostgreSQL, SQLite, DuckDB, MySQL: Accept keyword-like base window names in `OVER` [datafusion-sqlparser-rs]
via GitHub
-
[PR] SQLite: Parse the `BETWEEN` lower bound up to its `AND` [datafusion-sqlparser-rs]
via GitHub
-
[PR] MsSql: Keep the comma separator when displaying DECLARE lists [datafusion-sqlparser-rs]
via GitHub
-
Re: [PR] feat: support the S3A magic committer in native Parquet writes [datafusion-comet]
via GitHub
-
[PR] build(deps): bump tokio from 1.53.2 to 1.53.3 [datafusion-python]
via GitHub