Commit5a182b3

and

authored

BUG: fix empty suffix and prefix handling in pyarrow string methods (pandas-dev#63395)

Co-authored-by: Joris Van den Bossche <jorisvandenbossche@gmail.com>

1 parent0e978b6 commit5a182b3Copy full SHA for 5a182b3

File tree

+16

-2

lines changed

+16

-2

lines changed

Lines changed: 4 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -203,12 +203,16 @@ def _str_swapcase(self) -> Self:`
`203`	`203`	`returnself._from_pyarrow_array(pc.utf8_swapcase(self._pa_array))`
`204`	`204`
`205`	`205`	`def_str_removeprefix(self,prefix:str):`
	`206`	`+ifprefix=="":`
	`207`	`+returnself._from_pyarrow_array(self._pa_array)`
`206`	`208`	`starts_with=pc.starts_with(self._pa_array,pattern=prefix)`
`207`	`209`	`removed=pc.utf8_slice_codeunits(self._pa_array,len(prefix))`
`208`	`210`	`result=pc.if_else(starts_with,removed,self._pa_array)`
`209`	`211`	`returnself._from_pyarrow_array(result)`
`210`	`212`
`211`	`213`	`def_str_removesuffix(self,suffix:str):`
	`214`	`+ifsuffix=="":`
	`215`	`+returnself._from_pyarrow_array(self._pa_array)`
`212`	`216`	`ends_with=pc.ends_with(self._pa_array,pattern=suffix)`
`213`	`217`	`removed=pc.utf8_slice_codeunits(self._pa_array,0,stop=-len(suffix))`
`214`	`218`	`result=pc.if_else(ends_with,removed,self._pa_array)`

Lines changed: 12 additions & 2 deletions

Original file line number	Diff line number	Diff line change
`@@ -544,7 +544,12 @@ def test_strip_lstrip_rstrip_args(any_string_dtype, method, exp):`
`544`	`544`
`545`	`545`
`546`	`546`	`@pytest.mark.parametrize(`
`547`		`-"prefix, expected", [("a", ["b"," b c","bc"]), ("ab", ["","a b c","bc"])]`
	`547`	`+"prefix, expected",`
	`548`	`+ [`
	`549`	`+ ("a", ["b"," b c","bc"]),`
	`550`	`+ ("ab", ["","a b c","bc"]),`
	`551`	`+ ("", ["ab","a b c","bc"]),`
	`552`	`+ ],`
`548`	`553`	`)`
`549`	`554`	`deftest_removeprefix(any_string_dtype,prefix,expected):`
`550`	`555`	`ser=Series(["ab","a b c","bc"],dtype=any_string_dtype)`
`@@ -554,7 +559,12 @@ def test_removeprefix(any_string_dtype, prefix, expected):`
`554`	`559`
`555`	`560`
`556`	`561`	`@pytest.mark.parametrize(`
`557`		`-"suffix, expected", [("c", ["ab","a b ","b"]), ("bc", ["ab","a b c",""])]`
	`562`	`+"suffix, expected",`
	`563`	`+ [`
	`564`	`+ ("c", ["ab","a b ","b"]),`
	`565`	`+ ("bc", ["ab","a b c",""]),`
	`566`	`+ ("", ["ab","a b c","bc"]),`
	`567`	`+ ],`
`558`	`568`	`)`
`559`	`569`	`deftest_removesuffix(any_string_dtype,suffix,expected):`
`560`	`570`	`ser=Series(["ab","a b c","bc"],dtype=any_string_dtype)`

Comments

(0)