Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ENH: Enable pytables to round-trip with StringDtype #60663

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
Open

ENH: Enable pytables to round-trip with StringDtype #60663

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
91c8699
ENH: Enable pytables to round-trip with StringDtype
rhshadrach Jan 5, 2025
621ea78
Cleanups
rhshadrach Jan 5, 2025
11f8c0d
Remove type-ignore
rhshadrach Jan 5, 2025
52dbb5e
Add whatsnew note
rhshadrach Jan 8, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
54 changes: 48 additions & 6 deletions pandas/io/pytables.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,7 @@
writers as libwriters,
)
from pandas._libs.lib import is_string_array
from pandas._libs.missing import NA
from pandas._libs.tslibs import timezones
from pandas.compat._optional import import_optional_dependency
from pandas.compat.pickle_compat import patch_pickle
Expand Down Expand Up @@ -86,12 +87,16 @@
PeriodArray,
)
from pandas.core.arrays.datetimes import tz_to_dtype
from pandas.core.arrays.string_ import BaseStringArray
import pandas.core.common as com
from pandas.core.computation.pytables import (
PyTablesExpr,
maybe_expression,
)
from pandas.core.construction import extract_array
from pandas.core.construction import (
array as pd_array,
extract_array,
)
from pandas.core.indexes.api import ensure_index

from pandas.io.common import stringify_path
Expand Down Expand Up @@ -3023,6 +3028,18 @@ def read_array(self, key: str, start: int | None = None, stop: int | None = None

if isinstance(node, tables.VLArray):
ret = node[0][start:stop]
dtype = getattr(attrs, "value_type", None)
if dtype is not None:
if dtype == "str[python]":
dtype = StringDtype("python", np.nan)
elif dtype == "string[python]":
dtype = StringDtype("python", NA)
elif dtype == "str[pyarrow]":
dtype = StringDtype("pyarrow", np.nan)
else:
assert dtype == "string[pyarrow]"
dtype = StringDtype("pyarrow", NA)
Comment on lines +3033 to +3041
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure if there is a better approach here.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

StringDtype.construct_from_string?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This wouldn't allow round-tripping if you e.g. write out a Python-backed string with NaN-semantics, and read it in an environment with PyArrow installed.

ret = pd_array(ret, dtype=dtype)
else:
dtype = getattr(attrs, "value_type", None)
shape = getattr(attrs, "shape", None)
Expand Down Expand Up @@ -3262,6 +3279,19 @@ def write_array(
elif lib.is_np_dtype(value.dtype, "m"):
self._handle.create_array(self.group, key, value.view("i8"))
getattr(self.group, key)._v_attrs.value_type = "timedelta64"
elif isinstance(value, BaseStringArray):
vlarr = self._handle.create_vlarray(self.group, key, _tables().ObjectAtom())
vlarr.append(value.to_numpy())
node = getattr(self.group, key)
if value.dtype == StringDtype("python", np.nan):
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need all the branches here or can you just do str(values.dtype)?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Same as #60663 (comment)

node._v_attrs.value_type = "str[python]"
elif value.dtype == StringDtype("python", NA):
node._v_attrs.value_type = "string[python]"
elif value.dtype == StringDtype("pyarrow", np.nan):
node._v_attrs.value_type = "str[pyarrow]"
else:
assert value.dtype == StringDtype("pyarrow", NA)
node._v_attrs.value_type = "string[pyarrow]"
elif empty_array:
self.write_array_empty(key, value)
else:
Expand Down Expand Up @@ -3294,7 +3324,11 @@ def read(
index = self.read_index("index", start=start, stop=stop)
values = self.read_array("values", start=start, stop=stop)
result = Series(values, index=index, name=self.name, copy=False)
if using_string_dtype() and is_string_array(values, skipna=True):
if (
using_string_dtype()
and isinstance(values, np.ndarray)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm assuming read_array can return a non-ndarray type here? Does this prevent against pyarrow arrays?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm assuming read_array can return a non-ndarray type here?

Correct - datetime and string currently.

and is_string_array(values, skipna=True)
):
result = result.astype(StringDtype(na_value=np.nan))
return result

Expand Down Expand Up @@ -3363,7 +3397,11 @@ def read(

columns = items[items.get_indexer(blk_items)]
df = DataFrame(values.T, columns=columns, index=axes[1], copy=False)
if using_string_dtype() and is_string_array(values, skipna=True):
if (
using_string_dtype()
and isinstance(values, np.ndarray)
and is_string_array(values, skipna=True)
):
df = df.astype(StringDtype(na_value=np.nan))
dfs.append(df)

Expand Down Expand Up @@ -4737,9 +4775,13 @@ def read(
df = DataFrame._from_arrays([values], columns=cols_, index=index_)
if not (using_string_dtype() and values.dtype.kind == "O"):
assert (df.dtypes == values.dtype).all(), (df.dtypes, values.dtype)
if using_string_dtype() and is_string_array(
values, # type: ignore[arg-type]
skipna=True,
if (
using_string_dtype()
and isinstance(values, np.ndarray)
and is_string_array(
values, # type: ignore[arg-type]
skipna=True,
)
):
df = df.astype(StringDtype(na_value=np.nan))
frames.append(df)
Expand Down
68 changes: 54 additions & 14 deletions pandas/tests/io/pytables/test_put.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,8 +3,6 @@
import numpy as np
import pytest

from pandas._config import using_string_dtype

from pandas._libs.tslibs import Timestamp

import pandas as pd
Expand All @@ -26,7 +24,6 @@

pytestmark = [
pytest.mark.single_cpu,
pytest.mark.xfail(using_string_dtype(), reason="TODO(infer_string)", strict=False),
]


Expand Down Expand Up @@ -54,8 +51,8 @@ def test_api_default_format(tmp_path, setup_path):
with ensure_clean_store(setup_path) as store:
df = DataFrame(
1.1 * np.arange(120).reshape((30, 4)),
columns=Index(list("ABCD"), dtype=object),
index=Index([f"i-{i}" for i in range(30)], dtype=object),
columns=Index(list("ABCD")),
index=Index([f"i-{i}" for i in range(30)]),
)

with pd.option_context("io.hdf.default_format", "fixed"):
Expand All @@ -79,8 +76,8 @@ def test_api_default_format(tmp_path, setup_path):
path = tmp_path / setup_path
df = DataFrame(
1.1 * np.arange(120).reshape((30, 4)),
columns=Index(list("ABCD"), dtype=object),
index=Index([f"i-{i}" for i in range(30)], dtype=object),
columns=Index(list("ABCD")),
index=Index([f"i-{i}" for i in range(30)]),
)

with pd.option_context("io.hdf.default_format", "fixed"):
Expand All @@ -106,7 +103,7 @@ def test_put(setup_path):
)
df = DataFrame(
np.random.default_rng(2).standard_normal((20, 4)),
columns=Index(list("ABCD"), dtype=object),
columns=Index(list("ABCD")),
index=date_range("2000-01-01", periods=20, freq="B"),
)
store["a"] = ts
Expand Down Expand Up @@ -166,7 +163,7 @@ def test_put_compression(setup_path):
with ensure_clean_store(setup_path) as store:
df = DataFrame(
np.random.default_rng(2).standard_normal((10, 4)),
columns=Index(list("ABCD"), dtype=object),
columns=Index(list("ABCD")),
index=date_range("2000-01-01", periods=10, freq="B"),
)

Expand All @@ -183,7 +180,7 @@ def test_put_compression(setup_path):
def test_put_compression_blosc(setup_path):
df = DataFrame(
np.random.default_rng(2).standard_normal((10, 4)),
columns=Index(list("ABCD"), dtype=object),
columns=Index(list("ABCD")),
index=date_range("2000-01-01", periods=10, freq="B"),
)

Expand All @@ -197,10 +194,20 @@ def test_put_compression_blosc(setup_path):
tm.assert_frame_equal(store["c"], df)


def test_put_mixed_type(setup_path, performance_warning):
def test_put_datetime_ser(setup_path, performance_warning, using_infer_string):
# https://github.com/pandas-dev/pandas/pull/60663
ser = Series(3 * [Timestamp("20010102").as_unit("ns")])
with ensure_clean_store(setup_path) as store:
store.put("ser", ser)
expected = ser.copy()
result = store.get("ser")
tm.assert_series_equal(result, expected)


def test_put_mixed_type(setup_path, performance_warning, using_infer_string):
df = DataFrame(
np.random.default_rng(2).standard_normal((10, 4)),
columns=Index(list("ABCD"), dtype=object),
columns=Index(list("ABCD")),
index=date_range("2000-01-01", periods=10, freq="B"),
)
df["obj1"] = "foo"
Expand All @@ -220,13 +227,40 @@ def test_put_mixed_type(setup_path, performance_warning):
with ensure_clean_store(setup_path) as store:
_maybe_remove(store, "df")

with tm.assert_produces_warning(performance_warning):
warning = None if using_infer_string else performance_warning
with tm.assert_produces_warning(warning):
store.put("df", df)

expected = store.get("df")
tm.assert_frame_equal(expected, df)


def test_put_str_frame(setup_path, performance_warning, string_dtype_arguments):
# https://github.com/pandas-dev/pandas/pull/60663
dtype = pd.StringDtype(*string_dtype_arguments)
df = DataFrame({"a": pd.array(["x", pd.NA, "y"], dtype=dtype)})
with ensure_clean_store(setup_path) as store:
_maybe_remove(store, "df")

store.put("df", df)
expected = df
result = store.get("df")
tm.assert_frame_equal(result, expected)


def test_put_str_series(setup_path, performance_warning, string_dtype_arguments):
# https://github.com/pandas-dev/pandas/pull/60663
dtype = pd.StringDtype(*string_dtype_arguments)
ser = Series(["x", pd.NA, "y"], dtype=dtype)
with ensure_clean_store(setup_path) as store:
_maybe_remove(store, "df")

store.put("ser", ser)
expected = ser
result = store.get("ser")
tm.assert_series_equal(result, expected)


@pytest.mark.parametrize("format", ["table", "fixed"])
@pytest.mark.parametrize(
"index",
Expand All @@ -253,7 +287,7 @@ def test_store_index_types(setup_path, format, index):
tm.assert_frame_equal(df, store["df"])


def test_column_multiindex(setup_path):
def test_column_multiindex(setup_path, using_infer_string):
# GH 4710
# recreate multi-indexes properly

Expand All @@ -264,6 +298,12 @@ def test_column_multiindex(setup_path):
expected = df.set_axis(df.index.to_numpy())

with ensure_clean_store(setup_path) as store:
if using_infer_string:
# TODO(infer_string) make this work for string dtype
msg = "Saving a MultiIndex with an extension dtype is not supported."
with pytest.raises(NotImplementedError, match=msg):
store.put("df", df)
return
store.put("df", df)
tm.assert_frame_equal(
store["df"], expected, check_index_type=True, check_column_type=True
Expand Down
Loading