Commitb95f65b

authored

BUG: Fix regression in to_hdf with MultiIndex and StringDtype (GH#63412) (pandas-dev#63414)

1 parent9694eb1 commitb95f65bCopy full SHA for b95f65b

File tree

+13

-19

lines changed

+13

-19

lines changed

Lines changed: 3 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -3147,7 +3147,9 @@ def write_multi_index(self, key: str, index: MultiIndex) -> None:`
`3147`	`3147`	`zip(index.levels,index.codes,index.names,strict=True)`
`3148`	`3148`	`):`
`3149`	`3149`	`# write the level`
`3150`		`-ifisinstance(lev.dtype,ExtensionDtype):`
	`3150`	`+ifisinstance(lev.dtype,ExtensionDtype)andnotisinstance(`
	`3151`	`+lev.dtype,StringDtype`
	`3152`	`+ ):`
`3151`	`3153`	`raiseNotImplementedError(`
`3152`	`3154`	`"Saving a MultiIndex with an extension dtype is not supported."`
`3153`	`3155`	`)`

Lines changed: 0 additions & 6 deletions

Original file line number	Diff line number	Diff line change
`@@ -314,12 +314,6 @@ def test_column_multiindex(tmp_path, setup_path, using_infer_string):`
`314`	`314`
`315`	`315`	`path=tmp_path/setup_path`
`316`	`316`	`withHDFStore(path)asstore:`
`317`		`-ifusing_infer_string:`
`318`		`-# TODO(infer_string) make this work for string dtype`
`319`		`-msg="Saving a MultiIndex with an extension dtype is not supported."`
`320`		`-withpytest.raises(NotImplementedError,match=msg):`
`321`		`-store.put("df",df)`
`322`		`-return`
`323`	`317`	`store.put("df",df)`
`324`	`318`	`tm.assert_frame_equal(`
`325`	`319`	`store["df"],expected,check_index_type=True,check_column_type=True`

Lines changed: 0 additions & 6 deletions

Original file line number	Diff line number	Diff line change
`@@ -181,12 +181,6 @@ def test_read_hdf_open_store(tmp_path, setup_path, using_infer_string):`
`181`	`181`	`df=df.set_index(keys="E",append=True)`
`182`	`182`
`183`	`183`	`path=tmp_path/setup_path`
`184`		`-ifusing_infer_string:`
`185`		`-# TODO(infer_string) make this work for string dtype`
`186`		`-msg="Saving a MultiIndex with an extension dtype is not supported."`
`187`		`-withpytest.raises(NotImplementedError,match=msg):`
`188`		`-df.to_hdf(path,key="df",mode="w")`
`189`		`-return`
`190`	`184`	`df.to_hdf(path,key="df",mode="w")`
`191`	`185`	`direct=read_hdf(path,"df")`
`192`	`186`	`withHDFStore(path,mode="r")asstore:`

Lines changed: 0 additions & 6 deletions

Original file line number	Diff line number	Diff line change
`@@ -434,12 +434,6 @@ def test_store_hierarchical(`
`434`	`434`	`):`
`435`	`435`	`frame=multiindex_dataframe_random_data`
`436`	`436`
`437`		`-ifusing_infer_string:`
`438`		`-# TODO(infer_string) make this work for string dtype`
`439`		`-msg="Saving a MultiIndex with an extension dtype is not supported."`
`440`		`-withpytest.raises(NotImplementedError,match=msg):`
`441`		`-_check_roundtrip(frame,tm.assert_frame_equal,path=temp_file)`
`442`		`-return`
`443`	`437`	`_check_roundtrip(frame,tm.assert_frame_equal,path=temp_file)`
`444`	`438`	`_check_roundtrip(frame.T,tm.assert_frame_equal,path=temp_file)`
`445`	`439`	`_check_roundtrip(frame["A"],tm.assert_series_equal,path=temp_file)`

Lines changed: 10 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -1129,3 +1129,13 @@ def test_select_categorical_string_columns(tmp_path, model):`
`1129`	`1129`	`result=store.select("df","modelId == model")`
`1130`	`1130`	`expected=df[df["modelId"]==model]`
`1131`	`1131`	`tm.assert_frame_equal(result,expected)`
	`1132`	`+`
	`1133`	`+`
	`1134`	`+deftest_to_hdf_multiindex_string_dtype_crash(tmp_path):`
	`1135`	`+# GH#63412`
	`1136`	`+path=tmp_path/"test.h5"`
	`1137`	`+index=MultiIndex.from_tuples([("a","x"), ("b","y")],names=["level1","level2"])`
	`1138`	`+df=DataFrame({"value": [1,2]},index=index)`
	`1139`	`+df.to_hdf(path,key="test")`
	`1140`	`+result=read_hdf(path,key="test")`
	`1141`	`+tm.assert_frame_equal(df,result,check_dtype=False)`

Comments

(0)