Issues

pandas DataFrame operands: datastore.merge() runs eagerly, union() keeps a reference, join() raises AttributeError

chdb-io/chdb#773 · opened

on Nov 11, 2026

A user column named __orig_row_num__ breaks assign/filter/tail on a CSV source

chdb-io/chdb#772 · opened

on Nov 11, 2026

assert_datastore_equals_pandas() ignores check_index=True

chdb-io/chdb#771 · opened

on Nov 11, 2026

read_csv() and from_file() read exponent-notation floats such as 3.2e-05 as strings

chdb-io/chdb#770 · opened

on Nov 11, 2026

iloc[] executes the pipeline eagerly and returns a pandas DataFrame

chdb-io/chdb#769 · opened

on Nov 11, 2026

select() with aliases that swap two column names returns the columns unswapped

chdb-io/chdb#768 · opened

on Nov 11, 2026

where() with a condition from the unfiltered DataStore after a filter returns an extra row on a DataFrame source

chdb-io/chdb#767 · opened

on Nov 11, 2026

Arithmetic in assign() and column assignment widens small integer and float32 dtypes, unlike pandas

chdb-io/chdb#766 · opened

on Nov 11, 2026

Column assignment and mask() on a groupby().agg() result give wrong results or fail

chdb-io/chdb#765 · opened

on Nov 11, 2026

1

groupby(key)[col].agg('sum') returns a DataFrame instead of a Series

chdb-io/chdb#764 · opened

on Nov 11, 2026

skipna=False is ignored by groupby sum(), mean() and max()

chdb-io/chdb#763 · opened

on Nov 11, 2026

groupby() on a categorical column sorts groups by value instead of category order and loses the CategoricalIndex

chdb-io/chdb#762 · opened

on Nov 11, 2026

Row labels reset to 0..n-1 after groupby(as_index=False).agg() followed by sort_values() or a filter

chdb-io/chdb#761 · opened

on Nov 11, 2026

Two consecutive sort_values() followed by head() keep only the last sort

chdb-io/chdb#760 · opened

on Nov 11, 2026

groupby()[col].cumsum() on a file source returns values in group order instead of row order

chdb-io/chdb#759 · opened

on Nov 11, 2026

merge() on large CSV files returns rows in a different order from pandas, varying between runs

chdb-io/chdb#758 · opened

on Nov 11, 2026

Left merge on file sources returns nullable Int64 for unmatched integer columns where pandas returns float64

chdb-io/chdb#757 · opened

on Nov 11, 2026

merge() on file sources: assigned left columns come after the right columns, and left_on/right_on with the same name adds an r.k column

chdb-io/chdb#756 · opened

on Nov 11, 2026

join(on=left.id == right.id) on file sources compiles to ON "id" = "id" and returns a cross join

chdb-io/chdb#755 · opened

on Nov 11, 2026

join() on file sources applies the left side's head() and column selection after the join

chdb-io/chdb#754 · opened

on Nov 11, 2026

drop_duplicates() on a file source is not applied at its position in the chain

chdb-io/chdb#753 · opened

on Nov 11, 2026

1

drop_duplicates() with a single column name as subset fails on a file source

chdb-io/chdb#752 · opened

on Nov 11, 2026

DataStore and pandas disagree on Decimal columns; DB-API Decimal scale; Int128/Int256 as float64

chdb-io/chdb#745 · opened

on Nov 11, 2026

Working-directory Filesystem database: remaining gaps after #677

chdb-io/chdb#744 · opened

on Nov 11, 2026

Startup options of joining connections: remaining gaps after #676

chdb-io/chdb#743 · opened

on Nov 11, 2026

Navigation Menu

Repository navigation

Issues

pandas DataFrame operands: datastore.merge() runs eagerly, union() keeps a reference, join() raises AttributeError

A user column named __orig_row_num__ breaks assign/filter/tail on a CSV source

assert_datastore_equals_pandas() ignores check_index=True

read_csv() and from_file() read exponent-notation floats such as 3.2e-05 as strings

iloc[] executes the pipeline eagerly and returns a pandas DataFrame

select() with aliases that swap two column names returns the columns unswapped

where() with a condition from the unfiltered DataStore after a filter returns an extra row on a DataFrame source

Arithmetic in assign() and column assignment widens small integer and float32 dtypes, unlike pandas

Column assignment and mask() on a groupby().agg() result give wrong results or fail

groupby(key)[col].agg('sum') returns a DataFrame instead of a Series

skipna=False is ignored by groupby sum(), mean() and max()

groupby() on a categorical column sorts groups by value instead of category order and loses the CategoricalIndex

Row labels reset to 0..n-1 after groupby(as_index=False).agg() followed by sort_values() or a filter

Two consecutive sort_values() followed by head() keep only the last sort

groupby()[col].cumsum() on a file source returns values in group order instead of row order

merge() on large CSV files returns rows in a different order from pandas, varying between runs

Left merge on file sources returns nullable Int64 for unmatched integer columns where pandas returns float64

merge() on file sources: assigned left columns come after the right columns, and left_on/right_on with the same name adds an r.k column

join(on=left.id == right.id) on file sources compiles to ON "id" = "id" and returns a cross join

join() on file sources applies the left side's head() and column selection after the join

drop_duplicates() on a file source is not applied at its position in the chain

drop_duplicates() with a single column name as subset fails on a file source

DataStore and pandas disagree on Decimal columns; DB-API Decimal scale; Int128/Int256 as float64

Working-directory Filesystem database: remaining gaps after #677

Startup options of joining connections: remaining gaps after #676