duckdb/duckdb

564 open issues

• #24593 COUNT(x) rewritten to COUNT(*) doesn't prune x from the scan

  needs triage

  Investigate
• #24581 Chinese calendar: leap months skipped and new-moon boundary off by one day (357 days in 1900-2100)

  needs triage

  Investigate
• #24576 Orphaned temporary files are not cleaned up

  needs triage

  Investigate
• #24575 DuckDB Cli: duckbox lowercase column type rendering obscures actual type of ENUM-typed columns

  reproduced

  Investigate
• #24574 Named arguments (:=) bind correctly by name for macros, but silently fall back to positional binding for compiled/extension scalar functions

  Investigate
• #24570 DECIMAL syntax accepts non-integer parameters

  reproduced

  Investigate
• #24563 `HAVING` without `GROUP BY` in a **correlated** subquery is interpreted one way by `EXISTS` and the opposite way by the scalar form — the engine disagrees with itself about how many rows the subquery produced

  needs triage

  Investigate
• #24542 Wrong result: LIMIT+LEFT JOIN (v1.5 regression)

  reproduced

  Investigate
• #24541 Crash in CTE+LEFT JOIN query (v1.5 regression)

  needs triage

  Investigate
• #24540 Wrong result on LEFT JOIN ... WHERE col IS NULL (v1.5 regression)

  reproduced

  Investigate
• #24507 Parquet writer: data pages split only at a hardcoded 100 MB uncompressed (no option) — 100x larger pages than other writers, slow downstream reads

  Investigate
• #24481 MACRO overloads with DECIMAL-typed parameter aren't resolved

  reproduced

  Investigate
• #24474 Regression in DuckDB 1.5.5: HTTPFS cache no longer reused for repeated read_parquet() requests

  reproduced

  Investigate
• #24465 Vuln164: Calling dsdgen() twice with different scale factors and overwrite:=true corrupts the heap and causes a segfault

  needs triage

  Investigate
• #24464 Vuln163: dsdgen(sf := 'nan') bypasses scale-factor guards and aborts the process via a failed assertion in dsdgen-c

  needs triage

  Investigate
• #24462 Vuln162: Unbounded recursion in AnalyzeValueData/WriteValueData causes stack overflow on deeply nested VARIANT ARRAY values

  needs triage

  Investigate
• #24457 Vuln158: json_pretty() reported string length wraps around when the pretty-printed output exceeds UINT32_MAX bytes

  reproduced

  Investigate
• #24455 Row-dependent `APPROX_TOP_K` capacity is rejected by the combine window path

  reproduced

  Investigate
• #24454 Vuln151: Unbounded recursion in DeepMerge() causes native stack overflow and process crash on deeply nested JSON in json_deep_merge()

  needs triage

  Investigate
• #24453 Vuln150: ICU date_sub/date_diff throw an uncaught InternalException in SubtractField when isoyear/era parts span the BC/AD era boundary

  reproduced

  Investigate
• #24452 MODE(DISTINCT ...) can ignore the window ORDER BY input order

  reproduced

  Investigate
• #24451 HISTOGRAM_EXACT with row-dependent bins fails in a singleton window frame

  reproduced

  Investigate
• #24450 Vuln149: date_part with a non-foldable (column) part-name argument reaches InternalException when the runtime value names a DOUBLE-only part (epoch/julian)

  reproduced

  Investigate
• #24449 Top-N MIN with row-dependent n fails in a singleton window frame

  reproduced

  Investigate
• #24446 current_timestamp not listed in duckdb_functions()

  PR submitted

  Investigate
• #24444 Vuln142: to_aggregate_state() import of a crafted approx_top_k state with duplicate values triggers a reachable internal assertion in ApproxTopKImportState

  reproduced

  Investigate
• #24442 MAD FILTER over an excluded ROWS frame can include filtered/excluded rows

  reproduced

  Investigate
• #24441 Vuln139: unpivot_list bind function computes a merged common type via TryGetMaxLogicalType but never casts the bound arguments to it, causing a reachable INTERNAL Error at execution

  reproduced

  Investigate
• #24439 HISTOGRAM window combine double-counts NaN map keys

  reproduced

  Investigate
• #24431 Postgre Extension Error in R

  Investigate