Old engine for Continuous Time Bayesian Networks. Superseded by reCTBN. 🐍
https://github.com/madlabunimib/PyCTBN
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
1299 lines
44 KiB
1299 lines
44 KiB
from collections import OrderedDict
|
|
from datetime import date, datetime
|
|
import itertools
|
|
import operator
|
|
import re
|
|
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from pandas._libs.internals import BlockPlacement
|
|
|
|
import pandas as pd
|
|
from pandas import Categorical, DataFrame, DatetimeIndex, Index, MultiIndex, Series
|
|
import pandas._testing as tm
|
|
import pandas.core.algorithms as algos
|
|
from pandas.core.arrays import DatetimeArray, SparseArray, TimedeltaArray
|
|
from pandas.core.internals import BlockManager, SingleBlockManager, make_block
|
|
|
|
|
|
@pytest.fixture
|
|
def mgr():
|
|
return create_mgr(
|
|
"a: f8; b: object; c: f8; d: object; e: f8;"
|
|
"f: bool; g: i8; h: complex; i: datetime-1; j: datetime-2;"
|
|
"k: M8[ns, US/Eastern]; l: M8[ns, CET];"
|
|
)
|
|
|
|
|
|
def assert_block_equal(left, right):
|
|
tm.assert_numpy_array_equal(left.values, right.values)
|
|
assert left.dtype == right.dtype
|
|
assert isinstance(left.mgr_locs, BlockPlacement)
|
|
assert isinstance(right.mgr_locs, BlockPlacement)
|
|
tm.assert_numpy_array_equal(left.mgr_locs.as_array, right.mgr_locs.as_array)
|
|
|
|
|
|
def get_numeric_mat(shape):
|
|
arr = np.arange(shape[0])
|
|
return np.lib.stride_tricks.as_strided(
|
|
x=arr, shape=shape, strides=(arr.itemsize,) + (0,) * (len(shape) - 1)
|
|
).copy()
|
|
|
|
|
|
N = 10
|
|
|
|
|
|
def create_block(typestr, placement, item_shape=None, num_offset=0):
|
|
"""
|
|
Supported typestr:
|
|
|
|
* float, f8, f4, f2
|
|
* int, i8, i4, i2, i1
|
|
* uint, u8, u4, u2, u1
|
|
* complex, c16, c8
|
|
* bool
|
|
* object, string, O
|
|
* datetime, dt, M8[ns], M8[ns, tz]
|
|
* timedelta, td, m8[ns]
|
|
* sparse (SparseArray with fill_value=0.0)
|
|
* sparse_na (SparseArray with fill_value=np.nan)
|
|
* category, category2
|
|
|
|
"""
|
|
placement = BlockPlacement(placement)
|
|
num_items = len(placement)
|
|
|
|
if item_shape is None:
|
|
item_shape = (N,)
|
|
|
|
shape = (num_items,) + item_shape
|
|
|
|
mat = get_numeric_mat(shape)
|
|
|
|
if typestr in (
|
|
"float",
|
|
"f8",
|
|
"f4",
|
|
"f2",
|
|
"int",
|
|
"i8",
|
|
"i4",
|
|
"i2",
|
|
"i1",
|
|
"uint",
|
|
"u8",
|
|
"u4",
|
|
"u2",
|
|
"u1",
|
|
):
|
|
values = mat.astype(typestr) + num_offset
|
|
elif typestr in ("complex", "c16", "c8"):
|
|
values = 1.0j * (mat.astype(typestr) + num_offset)
|
|
elif typestr in ("object", "string", "O"):
|
|
values = np.reshape([f"A{i:d}" for i in mat.ravel() + num_offset], shape)
|
|
elif typestr in ("b", "bool"):
|
|
values = np.ones(shape, dtype=np.bool_)
|
|
elif typestr in ("datetime", "dt", "M8[ns]"):
|
|
values = (mat * 1e9).astype("M8[ns]")
|
|
elif typestr.startswith("M8[ns"):
|
|
# datetime with tz
|
|
m = re.search(r"M8\[ns,\s*(\w+\/?\w*)\]", typestr)
|
|
assert m is not None, f"incompatible typestr -> {typestr}"
|
|
tz = m.groups()[0]
|
|
assert num_items == 1, "must have only 1 num items for a tz-aware"
|
|
values = DatetimeIndex(np.arange(N) * 1e9, tz=tz)
|
|
elif typestr in ("timedelta", "td", "m8[ns]"):
|
|
values = (mat * 1).astype("m8[ns]")
|
|
elif typestr in ("category",):
|
|
values = Categorical([1, 1, 2, 2, 3, 3, 3, 3, 4, 4])
|
|
elif typestr in ("category2",):
|
|
values = Categorical(["a", "a", "a", "a", "b", "b", "c", "c", "c", "d"])
|
|
elif typestr in ("sparse", "sparse_na"):
|
|
# FIXME: doesn't support num_rows != 10
|
|
assert shape[-1] == 10
|
|
assert all(s == 1 for s in shape[:-1])
|
|
if typestr.endswith("_na"):
|
|
fill_value = np.nan
|
|
else:
|
|
fill_value = 0.0
|
|
values = SparseArray(
|
|
[fill_value, fill_value, 1, 2, 3, fill_value, 4, 5, fill_value, 6],
|
|
fill_value=fill_value,
|
|
)
|
|
arr = values.sp_values.view()
|
|
arr += num_offset - 1
|
|
else:
|
|
raise ValueError(f'Unsupported typestr: "{typestr}"')
|
|
|
|
return make_block(values, placement=placement, ndim=len(shape))
|
|
|
|
|
|
def create_single_mgr(typestr, num_rows=None):
|
|
if num_rows is None:
|
|
num_rows = N
|
|
|
|
return SingleBlockManager(
|
|
create_block(typestr, placement=slice(0, num_rows), item_shape=()),
|
|
Index(np.arange(num_rows)),
|
|
)
|
|
|
|
|
|
def create_mgr(descr, item_shape=None):
|
|
"""
|
|
Construct BlockManager from string description.
|
|
|
|
String description syntax looks similar to np.matrix initializer. It looks
|
|
like this::
|
|
|
|
a,b,c: f8; d,e,f: i8
|
|
|
|
Rules are rather simple:
|
|
|
|
* see list of supported datatypes in `create_block` method
|
|
* components are semicolon-separated
|
|
* each component is `NAME,NAME,NAME: DTYPE_ID`
|
|
* whitespace around colons & semicolons are removed
|
|
* components with same DTYPE_ID are combined into single block
|
|
* to force multiple blocks with same dtype, use '-SUFFIX'::
|
|
|
|
'a:f8-1; b:f8-2; c:f8-foobar'
|
|
|
|
"""
|
|
if item_shape is None:
|
|
item_shape = (N,)
|
|
|
|
offset = 0
|
|
mgr_items = []
|
|
block_placements = OrderedDict()
|
|
for d in descr.split(";"):
|
|
d = d.strip()
|
|
if not len(d):
|
|
continue
|
|
names, blockstr = d.partition(":")[::2]
|
|
blockstr = blockstr.strip()
|
|
names = names.strip().split(",")
|
|
|
|
mgr_items.extend(names)
|
|
placement = list(np.arange(len(names)) + offset)
|
|
try:
|
|
block_placements[blockstr].extend(placement)
|
|
except KeyError:
|
|
block_placements[blockstr] = placement
|
|
offset += len(names)
|
|
|
|
mgr_items = Index(mgr_items)
|
|
|
|
blocks = []
|
|
num_offset = 0
|
|
for blockstr, placement in block_placements.items():
|
|
typestr = blockstr.split("-")[0]
|
|
blocks.append(
|
|
create_block(
|
|
typestr, placement, item_shape=item_shape, num_offset=num_offset
|
|
)
|
|
)
|
|
num_offset += len(placement)
|
|
|
|
return BlockManager(
|
|
sorted(blocks, key=lambda b: b.mgr_locs[0]),
|
|
[mgr_items] + [np.arange(n) for n in item_shape],
|
|
)
|
|
|
|
|
|
class TestBlock:
|
|
def setup_method(self, method):
|
|
self.fblock = create_block("float", [0, 2, 4])
|
|
self.cblock = create_block("complex", [7])
|
|
self.oblock = create_block("object", [1, 3])
|
|
self.bool_block = create_block("bool", [5])
|
|
|
|
def test_constructor(self):
|
|
int32block = create_block("i4", [0])
|
|
assert int32block.dtype == np.int32
|
|
|
|
def test_pickle(self):
|
|
def _check(blk):
|
|
assert_block_equal(tm.round_trip_pickle(blk), blk)
|
|
|
|
_check(self.fblock)
|
|
_check(self.cblock)
|
|
_check(self.oblock)
|
|
_check(self.bool_block)
|
|
|
|
def test_mgr_locs(self):
|
|
assert isinstance(self.fblock.mgr_locs, BlockPlacement)
|
|
tm.assert_numpy_array_equal(
|
|
self.fblock.mgr_locs.as_array, np.array([0, 2, 4], dtype=np.int64)
|
|
)
|
|
|
|
def test_attrs(self):
|
|
assert self.fblock.shape == self.fblock.values.shape
|
|
assert self.fblock.dtype == self.fblock.values.dtype
|
|
assert len(self.fblock) == len(self.fblock.values)
|
|
|
|
def test_copy(self):
|
|
cop = self.fblock.copy()
|
|
assert cop is not self.fblock
|
|
assert_block_equal(self.fblock, cop)
|
|
|
|
def test_delete(self):
|
|
newb = self.fblock.copy()
|
|
newb.delete(0)
|
|
assert isinstance(newb.mgr_locs, BlockPlacement)
|
|
tm.assert_numpy_array_equal(
|
|
newb.mgr_locs.as_array, np.array([2, 4], dtype=np.int64)
|
|
)
|
|
assert (newb.values[0] == 1).all()
|
|
|
|
newb = self.fblock.copy()
|
|
newb.delete(1)
|
|
assert isinstance(newb.mgr_locs, BlockPlacement)
|
|
tm.assert_numpy_array_equal(
|
|
newb.mgr_locs.as_array, np.array([0, 4], dtype=np.int64)
|
|
)
|
|
assert (newb.values[1] == 2).all()
|
|
|
|
newb = self.fblock.copy()
|
|
newb.delete(2)
|
|
tm.assert_numpy_array_equal(
|
|
newb.mgr_locs.as_array, np.array([0, 2], dtype=np.int64)
|
|
)
|
|
assert (newb.values[1] == 1).all()
|
|
|
|
newb = self.fblock.copy()
|
|
|
|
with pytest.raises(IndexError, match=None):
|
|
newb.delete(3)
|
|
|
|
|
|
class TestBlockManager:
|
|
def test_attrs(self):
|
|
mgr = create_mgr("a,b,c: f8-1; d,e,f: f8-2")
|
|
assert mgr.nblocks == 2
|
|
assert len(mgr) == 6
|
|
|
|
def test_is_mixed_dtype(self):
|
|
assert not create_mgr("a,b:f8").is_mixed_type
|
|
assert not create_mgr("a:f8-1; b:f8-2").is_mixed_type
|
|
|
|
assert create_mgr("a,b:f8; c,d: f4").is_mixed_type
|
|
assert create_mgr("a,b:f8; c,d: object").is_mixed_type
|
|
|
|
def test_duplicate_ref_loc_failure(self):
|
|
tmp_mgr = create_mgr("a:bool; a: f8")
|
|
|
|
axes, blocks = tmp_mgr.axes, tmp_mgr.blocks
|
|
|
|
blocks[0].mgr_locs = np.array([0])
|
|
blocks[1].mgr_locs = np.array([0])
|
|
|
|
# test trying to create block manager with overlapping ref locs
|
|
|
|
msg = "Gaps in blk ref_locs"
|
|
|
|
with pytest.raises(AssertionError, match=msg):
|
|
mgr = BlockManager(blocks, axes)
|
|
mgr._rebuild_blknos_and_blklocs()
|
|
|
|
blocks[0].mgr_locs = np.array([0])
|
|
blocks[1].mgr_locs = np.array([1])
|
|
mgr = BlockManager(blocks, axes)
|
|
mgr.iget(1)
|
|
|
|
def test_pickle(self, mgr):
|
|
|
|
mgr2 = tm.round_trip_pickle(mgr)
|
|
tm.assert_frame_equal(DataFrame(mgr), DataFrame(mgr2))
|
|
|
|
# GH2431
|
|
assert hasattr(mgr2, "_is_consolidated")
|
|
assert hasattr(mgr2, "_known_consolidated")
|
|
|
|
# reset to False on load
|
|
assert not mgr2._is_consolidated
|
|
assert not mgr2._known_consolidated
|
|
|
|
@pytest.mark.parametrize("mgr_string", ["a,a,a:f8", "a: f8; a: i8"])
|
|
def test_non_unique_pickle(self, mgr_string):
|
|
mgr = create_mgr(mgr_string)
|
|
mgr2 = tm.round_trip_pickle(mgr)
|
|
tm.assert_frame_equal(DataFrame(mgr), DataFrame(mgr2))
|
|
|
|
def test_categorical_block_pickle(self):
|
|
mgr = create_mgr("a: category")
|
|
mgr2 = tm.round_trip_pickle(mgr)
|
|
tm.assert_frame_equal(DataFrame(mgr), DataFrame(mgr2))
|
|
|
|
smgr = create_single_mgr("category")
|
|
smgr2 = tm.round_trip_pickle(smgr)
|
|
tm.assert_series_equal(Series(smgr), Series(smgr2))
|
|
|
|
def test_iget(self):
|
|
cols = Index(list("abc"))
|
|
values = np.random.rand(3, 3)
|
|
block = make_block(values=values.copy(), placement=np.arange(3))
|
|
mgr = BlockManager(blocks=[block], axes=[cols, np.arange(3)])
|
|
|
|
tm.assert_almost_equal(mgr.iget(0).internal_values(), values[0])
|
|
tm.assert_almost_equal(mgr.iget(1).internal_values(), values[1])
|
|
tm.assert_almost_equal(mgr.iget(2).internal_values(), values[2])
|
|
|
|
def test_set(self):
|
|
mgr = create_mgr("a,b,c: int", item_shape=(3,))
|
|
|
|
mgr.insert(len(mgr.items), "d", np.array(["foo"] * 3))
|
|
mgr.iset(1, np.array(["bar"] * 3))
|
|
tm.assert_numpy_array_equal(mgr.iget(0).internal_values(), np.array([0] * 3))
|
|
tm.assert_numpy_array_equal(
|
|
mgr.iget(1).internal_values(), np.array(["bar"] * 3, dtype=np.object_)
|
|
)
|
|
tm.assert_numpy_array_equal(mgr.iget(2).internal_values(), np.array([2] * 3))
|
|
tm.assert_numpy_array_equal(
|
|
mgr.iget(3).internal_values(), np.array(["foo"] * 3, dtype=np.object_)
|
|
)
|
|
|
|
def test_set_change_dtype(self, mgr):
|
|
mgr.insert(len(mgr.items), "baz", np.zeros(N, dtype=bool))
|
|
|
|
mgr.iset(mgr.items.get_loc("baz"), np.repeat("foo", N))
|
|
idx = mgr.items.get_loc("baz")
|
|
assert mgr.iget(idx).dtype == np.object_
|
|
|
|
mgr2 = mgr.consolidate()
|
|
mgr2.iset(mgr2.items.get_loc("baz"), np.repeat("foo", N))
|
|
idx = mgr2.items.get_loc("baz")
|
|
assert mgr2.iget(idx).dtype == np.object_
|
|
|
|
mgr2.insert(len(mgr2.items), "quux", tm.randn(N).astype(int))
|
|
idx = mgr2.items.get_loc("quux")
|
|
assert mgr2.iget(idx).dtype == np.int_
|
|
|
|
mgr2.iset(mgr2.items.get_loc("quux"), tm.randn(N))
|
|
assert mgr2.iget(idx).dtype == np.float_
|
|
|
|
def test_copy(self, mgr):
|
|
cp = mgr.copy(deep=False)
|
|
for blk, cp_blk in zip(mgr.blocks, cp.blocks):
|
|
|
|
# view assertion
|
|
tm.assert_equal(cp_blk.values, blk.values)
|
|
if isinstance(blk.values, np.ndarray):
|
|
assert cp_blk.values.base is blk.values.base
|
|
else:
|
|
# DatetimeTZBlock has DatetimeIndex values
|
|
assert cp_blk.values._data.base is blk.values._data.base
|
|
|
|
cp = mgr.copy(deep=True)
|
|
for blk, cp_blk in zip(mgr.blocks, cp.blocks):
|
|
|
|
# copy assertion we either have a None for a base or in case of
|
|
# some blocks it is an array (e.g. datetimetz), but was copied
|
|
tm.assert_equal(cp_blk.values, blk.values)
|
|
if not isinstance(cp_blk.values, np.ndarray):
|
|
assert cp_blk.values._data.base is not blk.values._data.base
|
|
else:
|
|
assert cp_blk.values.base is None and blk.values.base is None
|
|
|
|
def test_sparse(self):
|
|
mgr = create_mgr("a: sparse-1; b: sparse-2")
|
|
# what to test here?
|
|
assert mgr.as_array().dtype == np.float64
|
|
|
|
def test_sparse_mixed(self):
|
|
mgr = create_mgr("a: sparse-1; b: sparse-2; c: f8")
|
|
assert len(mgr.blocks) == 3
|
|
assert isinstance(mgr, BlockManager)
|
|
|
|
# TODO: what to test here?
|
|
|
|
@pytest.mark.parametrize(
|
|
"mgr_string, dtype",
|
|
[("c: f4; d: f2", np.float32), ("c: f4; d: f2; e: f8", np.float64)],
|
|
)
|
|
def test_as_array_float(self, mgr_string, dtype):
|
|
mgr = create_mgr(mgr_string)
|
|
assert mgr.as_array().dtype == dtype
|
|
|
|
@pytest.mark.parametrize(
|
|
"mgr_string, dtype",
|
|
[
|
|
("a: bool-1; b: bool-2", np.bool_),
|
|
("a: i8-1; b: i8-2; c: i4; d: i2; e: u1", np.int64),
|
|
("c: i4; d: i2; e: u1", np.int32),
|
|
],
|
|
)
|
|
def test_as_array_int_bool(self, mgr_string, dtype):
|
|
mgr = create_mgr(mgr_string)
|
|
assert mgr.as_array().dtype == dtype
|
|
|
|
def test_as_array_datetime(self):
|
|
mgr = create_mgr("h: datetime-1; g: datetime-2")
|
|
assert mgr.as_array().dtype == "M8[ns]"
|
|
|
|
def test_as_array_datetime_tz(self):
|
|
mgr = create_mgr("h: M8[ns, US/Eastern]; g: M8[ns, CET]")
|
|
assert mgr.iget(0).dtype == "datetime64[ns, US/Eastern]"
|
|
assert mgr.iget(1).dtype == "datetime64[ns, CET]"
|
|
assert mgr.as_array().dtype == "object"
|
|
|
|
@pytest.mark.parametrize("t", ["float16", "float32", "float64", "int32", "int64"])
|
|
def test_astype(self, t):
|
|
# coerce all
|
|
mgr = create_mgr("c: f4; d: f2; e: f8")
|
|
|
|
t = np.dtype(t)
|
|
tmgr = mgr.astype(t)
|
|
assert tmgr.iget(0).dtype.type == t
|
|
assert tmgr.iget(1).dtype.type == t
|
|
assert tmgr.iget(2).dtype.type == t
|
|
|
|
# mixed
|
|
mgr = create_mgr("a,b: object; c: bool; d: datetime; e: f4; f: f2; g: f8")
|
|
|
|
t = np.dtype(t)
|
|
tmgr = mgr.astype(t, errors="ignore")
|
|
assert tmgr.iget(2).dtype.type == t
|
|
assert tmgr.iget(4).dtype.type == t
|
|
assert tmgr.iget(5).dtype.type == t
|
|
assert tmgr.iget(6).dtype.type == t
|
|
|
|
assert tmgr.iget(0).dtype.type == np.object_
|
|
assert tmgr.iget(1).dtype.type == np.object_
|
|
if t != np.int64:
|
|
assert tmgr.iget(3).dtype.type == np.datetime64
|
|
else:
|
|
assert tmgr.iget(3).dtype.type == t
|
|
|
|
def test_convert(self):
|
|
def _compare(old_mgr, new_mgr):
|
|
""" compare the blocks, numeric compare ==, object don't """
|
|
old_blocks = set(old_mgr.blocks)
|
|
new_blocks = set(new_mgr.blocks)
|
|
assert len(old_blocks) == len(new_blocks)
|
|
|
|
# compare non-numeric
|
|
for b in old_blocks:
|
|
found = False
|
|
for nb in new_blocks:
|
|
if (b.values == nb.values).all():
|
|
found = True
|
|
break
|
|
assert found
|
|
|
|
for b in new_blocks:
|
|
found = False
|
|
for ob in old_blocks:
|
|
if (b.values == ob.values).all():
|
|
found = True
|
|
break
|
|
assert found
|
|
|
|
# noops
|
|
mgr = create_mgr("f: i8; g: f8")
|
|
new_mgr = mgr.convert()
|
|
_compare(mgr, new_mgr)
|
|
|
|
# convert
|
|
mgr = create_mgr("a,b,foo: object; f: i8; g: f8")
|
|
mgr.iset(0, np.array(["1"] * N, dtype=np.object_))
|
|
mgr.iset(1, np.array(["2."] * N, dtype=np.object_))
|
|
mgr.iset(2, np.array(["foo."] * N, dtype=np.object_))
|
|
new_mgr = mgr.convert(numeric=True)
|
|
assert new_mgr.iget(0).dtype == np.int64
|
|
assert new_mgr.iget(1).dtype == np.float64
|
|
assert new_mgr.iget(2).dtype == np.object_
|
|
assert new_mgr.iget(3).dtype == np.int64
|
|
assert new_mgr.iget(4).dtype == np.float64
|
|
|
|
mgr = create_mgr(
|
|
"a,b,foo: object; f: i4; bool: bool; dt: datetime; i: i8; g: f8; h: f2"
|
|
)
|
|
mgr.iset(0, np.array(["1"] * N, dtype=np.object_))
|
|
mgr.iset(1, np.array(["2."] * N, dtype=np.object_))
|
|
mgr.iset(2, np.array(["foo."] * N, dtype=np.object_))
|
|
new_mgr = mgr.convert(numeric=True)
|
|
assert new_mgr.iget(0).dtype == np.int64
|
|
assert new_mgr.iget(1).dtype == np.float64
|
|
assert new_mgr.iget(2).dtype == np.object_
|
|
assert new_mgr.iget(3).dtype == np.int32
|
|
assert new_mgr.iget(4).dtype == np.bool_
|
|
assert new_mgr.iget(5).dtype.type, np.datetime64
|
|
assert new_mgr.iget(6).dtype == np.int64
|
|
assert new_mgr.iget(7).dtype == np.float64
|
|
assert new_mgr.iget(8).dtype == np.float16
|
|
|
|
def test_invalid_ea_block(self):
|
|
with pytest.raises(AssertionError, match="block.size != values.size"):
|
|
create_mgr("a: category; b: category")
|
|
|
|
with pytest.raises(AssertionError, match="block.size != values.size"):
|
|
create_mgr("a: category2; b: category2")
|
|
|
|
def test_interleave(self):
|
|
# self
|
|
for dtype in ["f8", "i8", "object", "bool", "complex", "M8[ns]", "m8[ns]"]:
|
|
mgr = create_mgr(f"a: {dtype}")
|
|
assert mgr.as_array().dtype == dtype
|
|
mgr = create_mgr(f"a: {dtype}; b: {dtype}")
|
|
assert mgr.as_array().dtype == dtype
|
|
|
|
@pytest.mark.parametrize(
|
|
"mgr_string, dtype",
|
|
[
|
|
("a: category", "i8"),
|
|
("a: category; b: category", "i8"),
|
|
("a: category; b: category2", "object"),
|
|
("a: category2", "object"),
|
|
("a: category2; b: category2", "object"),
|
|
("a: f8", "f8"),
|
|
("a: f8; b: i8", "f8"),
|
|
("a: f4; b: i8", "f8"),
|
|
("a: f4; b: i8; d: object", "object"),
|
|
("a: bool; b: i8", "object"),
|
|
("a: complex", "complex"),
|
|
("a: f8; b: category", "object"),
|
|
("a: M8[ns]; b: category", "object"),
|
|
("a: M8[ns]; b: bool", "object"),
|
|
("a: M8[ns]; b: i8", "object"),
|
|
("a: m8[ns]; b: bool", "object"),
|
|
("a: m8[ns]; b: i8", "object"),
|
|
("a: M8[ns]; b: m8[ns]", "object"),
|
|
],
|
|
)
|
|
def test_interleave_dtype(self, mgr_string, dtype):
|
|
# will be converted according the actual dtype of the underlying
|
|
mgr = create_mgr("a: category")
|
|
assert mgr.as_array().dtype == "i8"
|
|
mgr = create_mgr("a: category; b: category2")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: category2")
|
|
assert mgr.as_array().dtype == "object"
|
|
|
|
# combinations
|
|
mgr = create_mgr("a: f8")
|
|
assert mgr.as_array().dtype == "f8"
|
|
mgr = create_mgr("a: f8; b: i8")
|
|
assert mgr.as_array().dtype == "f8"
|
|
mgr = create_mgr("a: f4; b: i8")
|
|
assert mgr.as_array().dtype == "f8"
|
|
mgr = create_mgr("a: f4; b: i8; d: object")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: bool; b: i8")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: complex")
|
|
assert mgr.as_array().dtype == "complex"
|
|
mgr = create_mgr("a: f8; b: category")
|
|
assert mgr.as_array().dtype == "f8"
|
|
mgr = create_mgr("a: M8[ns]; b: category")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: M8[ns]; b: bool")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: M8[ns]; b: i8")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: m8[ns]; b: bool")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: m8[ns]; b: i8")
|
|
assert mgr.as_array().dtype == "object"
|
|
mgr = create_mgr("a: M8[ns]; b: m8[ns]")
|
|
assert mgr.as_array().dtype == "object"
|
|
|
|
def test_consolidate_ordering_issues(self, mgr):
|
|
mgr.iset(mgr.items.get_loc("f"), tm.randn(N))
|
|
mgr.iset(mgr.items.get_loc("d"), tm.randn(N))
|
|
mgr.iset(mgr.items.get_loc("b"), tm.randn(N))
|
|
mgr.iset(mgr.items.get_loc("g"), tm.randn(N))
|
|
mgr.iset(mgr.items.get_loc("h"), tm.randn(N))
|
|
|
|
# we have datetime/tz blocks in mgr
|
|
cons = mgr.consolidate()
|
|
assert cons.nblocks == 4
|
|
cons = mgr.consolidate().get_numeric_data()
|
|
assert cons.nblocks == 1
|
|
assert isinstance(cons.blocks[0].mgr_locs, BlockPlacement)
|
|
tm.assert_numpy_array_equal(
|
|
cons.blocks[0].mgr_locs.as_array, np.arange(len(cons.items), dtype=np.int64)
|
|
)
|
|
|
|
def test_reindex_items(self):
|
|
# mgr is not consolidated, f8 & f8-2 blocks
|
|
mgr = create_mgr("a: f8; b: i8; c: f8; d: i8; e: f8; f: bool; g: f8-2")
|
|
|
|
reindexed = mgr.reindex_axis(["g", "c", "a", "d"], axis=0)
|
|
assert reindexed.nblocks == 2
|
|
tm.assert_index_equal(reindexed.items, pd.Index(["g", "c", "a", "d"]))
|
|
tm.assert_almost_equal(
|
|
mgr.iget(6).internal_values(), reindexed.iget(0).internal_values()
|
|
)
|
|
tm.assert_almost_equal(
|
|
mgr.iget(2).internal_values(), reindexed.iget(1).internal_values()
|
|
)
|
|
tm.assert_almost_equal(
|
|
mgr.iget(0).internal_values(), reindexed.iget(2).internal_values()
|
|
)
|
|
tm.assert_almost_equal(
|
|
mgr.iget(3).internal_values(), reindexed.iget(3).internal_values()
|
|
)
|
|
|
|
def test_get_numeric_data(self):
|
|
mgr = create_mgr(
|
|
"int: int; float: float; complex: complex;"
|
|
"str: object; bool: bool; obj: object; dt: datetime",
|
|
item_shape=(3,),
|
|
)
|
|
mgr.iset(5, np.array([1, 2, 3], dtype=np.object_))
|
|
|
|
numeric = mgr.get_numeric_data()
|
|
tm.assert_index_equal(
|
|
numeric.items, pd.Index(["int", "float", "complex", "bool"])
|
|
)
|
|
tm.assert_almost_equal(
|
|
mgr.iget(mgr.items.get_loc("float")).internal_values(),
|
|
numeric.iget(numeric.items.get_loc("float")).internal_values(),
|
|
)
|
|
|
|
# Check sharing
|
|
numeric.iset(numeric.items.get_loc("float"), np.array([100.0, 200.0, 300.0]))
|
|
tm.assert_almost_equal(
|
|
mgr.iget(mgr.items.get_loc("float")).internal_values(),
|
|
np.array([100.0, 200.0, 300.0]),
|
|
)
|
|
|
|
numeric2 = mgr.get_numeric_data(copy=True)
|
|
tm.assert_index_equal(
|
|
numeric.items, pd.Index(["int", "float", "complex", "bool"])
|
|
)
|
|
numeric2.iset(
|
|
numeric2.items.get_loc("float"), np.array([1000.0, 2000.0, 3000.0])
|
|
)
|
|
tm.assert_almost_equal(
|
|
mgr.iget(mgr.items.get_loc("float")).internal_values(),
|
|
np.array([100.0, 200.0, 300.0]),
|
|
)
|
|
|
|
def test_get_bool_data(self):
|
|
mgr = create_mgr(
|
|
"int: int; float: float; complex: complex;"
|
|
"str: object; bool: bool; obj: object; dt: datetime",
|
|
item_shape=(3,),
|
|
)
|
|
mgr.iset(6, np.array([True, False, True], dtype=np.object_))
|
|
|
|
bools = mgr.get_bool_data()
|
|
tm.assert_index_equal(bools.items, pd.Index(["bool"]))
|
|
tm.assert_almost_equal(
|
|
mgr.iget(mgr.items.get_loc("bool")).internal_values(),
|
|
bools.iget(bools.items.get_loc("bool")).internal_values(),
|
|
)
|
|
|
|
bools.iset(0, np.array([True, False, True]))
|
|
tm.assert_numpy_array_equal(
|
|
mgr.iget(mgr.items.get_loc("bool")).internal_values(),
|
|
np.array([True, False, True]),
|
|
)
|
|
|
|
# Check sharing
|
|
bools2 = mgr.get_bool_data(copy=True)
|
|
bools2.iset(0, np.array([False, True, False]))
|
|
tm.assert_numpy_array_equal(
|
|
mgr.iget(mgr.items.get_loc("bool")).internal_values(),
|
|
np.array([True, False, True]),
|
|
)
|
|
|
|
def test_unicode_repr_doesnt_raise(self):
|
|
repr(create_mgr("b,\u05d0: object"))
|
|
|
|
@pytest.mark.parametrize(
|
|
"mgr_string", ["a,b,c: i8-1; d,e,f: i8-2", "a,a,a: i8-1; b,b,b: i8-2"]
|
|
)
|
|
def test_equals(self, mgr_string):
|
|
# unique items
|
|
bm1 = create_mgr(mgr_string)
|
|
bm2 = BlockManager(bm1.blocks[::-1], bm1.axes)
|
|
assert bm1.equals(bm2)
|
|
|
|
@pytest.mark.parametrize(
|
|
"mgr_string",
|
|
[
|
|
"a:i8;b:f8", # basic case
|
|
"a:i8;b:f8;c:c8;d:b", # many types
|
|
"a:i8;e:dt;f:td;g:string", # more types
|
|
"a:i8;b:category;c:category2", # categories
|
|
"c:sparse;d:sparse_na;b:f8", # sparse
|
|
],
|
|
)
|
|
def test_equals_block_order_different_dtypes(self, mgr_string):
|
|
# GH 9330
|
|
bm = create_mgr(mgr_string)
|
|
block_perms = itertools.permutations(bm.blocks)
|
|
for bm_perm in block_perms:
|
|
bm_this = BlockManager(bm_perm, bm.axes)
|
|
assert bm.equals(bm_this)
|
|
assert bm_this.equals(bm)
|
|
|
|
def test_single_mgr_ctor(self):
|
|
mgr = create_single_mgr("f8", num_rows=5)
|
|
assert mgr.as_array().tolist() == [0.0, 1.0, 2.0, 3.0, 4.0]
|
|
|
|
@pytest.mark.parametrize("value", [1, "True", [1, 2, 3], 5.0])
|
|
def test_validate_bool_args(self, value):
|
|
bm1 = create_mgr("a,b,c: i8-1; d,e,f: i8-2")
|
|
|
|
msg = (
|
|
'For argument "inplace" expected type bool, '
|
|
f"received type {type(value).__name__}."
|
|
)
|
|
with pytest.raises(ValueError, match=msg):
|
|
bm1.replace_list([1], [2], inplace=value)
|
|
|
|
|
|
class TestIndexing:
|
|
# Nosetests-style data-driven tests.
|
|
#
|
|
# This test applies different indexing routines to block managers and
|
|
# compares the outcome to the result of same operations on np.ndarray.
|
|
#
|
|
# NOTE: sparse (SparseBlock with fill_value != np.nan) fail a lot of tests
|
|
# and are disabled.
|
|
|
|
MANAGERS = [
|
|
create_single_mgr("f8", N),
|
|
create_single_mgr("i8", N),
|
|
# 2-dim
|
|
create_mgr("a,b,c,d,e,f: f8", item_shape=(N,)),
|
|
create_mgr("a,b,c,d,e,f: i8", item_shape=(N,)),
|
|
create_mgr("a,b: f8; c,d: i8; e,f: string", item_shape=(N,)),
|
|
create_mgr("a,b: f8; c,d: i8; e,f: f8", item_shape=(N,)),
|
|
]
|
|
|
|
@pytest.mark.parametrize("mgr", MANAGERS)
|
|
def test_get_slice(self, mgr):
|
|
def assert_slice_ok(mgr, axis, slobj):
|
|
mat = mgr.as_array()
|
|
|
|
# we maybe using an ndarray to test slicing and
|
|
# might not be the full length of the axis
|
|
if isinstance(slobj, np.ndarray):
|
|
ax = mgr.axes[axis]
|
|
if len(ax) and len(slobj) and len(slobj) != len(ax):
|
|
slobj = np.concatenate(
|
|
[slobj, np.zeros(len(ax) - len(slobj), dtype=bool)]
|
|
)
|
|
sliced = mgr.get_slice(slobj, axis=axis)
|
|
mat_slobj = (slice(None),) * axis + (slobj,)
|
|
tm.assert_numpy_array_equal(
|
|
mat[mat_slobj], sliced.as_array(), check_dtype=False
|
|
)
|
|
tm.assert_index_equal(mgr.axes[axis][slobj], sliced.axes[axis])
|
|
|
|
assert mgr.ndim <= 2, mgr.ndim
|
|
for ax in range(mgr.ndim):
|
|
# slice
|
|
assert_slice_ok(mgr, ax, slice(None))
|
|
assert_slice_ok(mgr, ax, slice(3))
|
|
assert_slice_ok(mgr, ax, slice(100))
|
|
assert_slice_ok(mgr, ax, slice(1, 4))
|
|
assert_slice_ok(mgr, ax, slice(3, 0, -2))
|
|
|
|
# boolean mask
|
|
assert_slice_ok(mgr, ax, np.array([], dtype=np.bool_))
|
|
assert_slice_ok(mgr, ax, np.ones(mgr.shape[ax], dtype=np.bool_))
|
|
assert_slice_ok(mgr, ax, np.zeros(mgr.shape[ax], dtype=np.bool_))
|
|
|
|
if mgr.shape[ax] >= 3:
|
|
assert_slice_ok(mgr, ax, np.arange(mgr.shape[ax]) % 3 == 0)
|
|
assert_slice_ok(mgr, ax, np.array([True, True, False], dtype=np.bool_))
|
|
|
|
# fancy indexer
|
|
assert_slice_ok(mgr, ax, [])
|
|
assert_slice_ok(mgr, ax, list(range(mgr.shape[ax])))
|
|
|
|
if mgr.shape[ax] >= 3:
|
|
assert_slice_ok(mgr, ax, [0, 1, 2])
|
|
assert_slice_ok(mgr, ax, [-1, -2, -3])
|
|
|
|
@pytest.mark.parametrize("mgr", MANAGERS)
|
|
def test_take(self, mgr):
|
|
def assert_take_ok(mgr, axis, indexer):
|
|
mat = mgr.as_array()
|
|
taken = mgr.take(indexer, axis)
|
|
tm.assert_numpy_array_equal(
|
|
np.take(mat, indexer, axis), taken.as_array(), check_dtype=False
|
|
)
|
|
tm.assert_index_equal(mgr.axes[axis].take(indexer), taken.axes[axis])
|
|
|
|
for ax in range(mgr.ndim):
|
|
# take/fancy indexer
|
|
assert_take_ok(mgr, ax, indexer=[])
|
|
assert_take_ok(mgr, ax, indexer=[0, 0, 0])
|
|
assert_take_ok(mgr, ax, indexer=list(range(mgr.shape[ax])))
|
|
|
|
if mgr.shape[ax] >= 3:
|
|
assert_take_ok(mgr, ax, indexer=[0, 1, 2])
|
|
assert_take_ok(mgr, ax, indexer=[-1, -2, -3])
|
|
|
|
@pytest.mark.parametrize("mgr", MANAGERS)
|
|
@pytest.mark.parametrize("fill_value", [None, np.nan, 100.0])
|
|
def test_reindex_axis(self, fill_value, mgr):
|
|
def assert_reindex_axis_is_ok(mgr, axis, new_labels, fill_value):
|
|
mat = mgr.as_array()
|
|
indexer = mgr.axes[axis].get_indexer_for(new_labels)
|
|
|
|
reindexed = mgr.reindex_axis(new_labels, axis, fill_value=fill_value)
|
|
tm.assert_numpy_array_equal(
|
|
algos.take_nd(mat, indexer, axis, fill_value=fill_value),
|
|
reindexed.as_array(),
|
|
check_dtype=False,
|
|
)
|
|
tm.assert_index_equal(reindexed.axes[axis], new_labels)
|
|
|
|
for ax in range(mgr.ndim):
|
|
assert_reindex_axis_is_ok(mgr, ax, pd.Index([]), fill_value)
|
|
assert_reindex_axis_is_ok(mgr, ax, mgr.axes[ax], fill_value)
|
|
assert_reindex_axis_is_ok(mgr, ax, mgr.axes[ax][[0, 0, 0]], fill_value)
|
|
assert_reindex_axis_is_ok(
|
|
mgr, ax, pd.Index(["foo", "bar", "baz"]), fill_value
|
|
)
|
|
assert_reindex_axis_is_ok(
|
|
mgr, ax, pd.Index(["foo", mgr.axes[ax][0], "baz"]), fill_value
|
|
)
|
|
|
|
if mgr.shape[ax] >= 3:
|
|
assert_reindex_axis_is_ok(mgr, ax, mgr.axes[ax][:-3], fill_value)
|
|
assert_reindex_axis_is_ok(mgr, ax, mgr.axes[ax][-3::-1], fill_value)
|
|
assert_reindex_axis_is_ok(
|
|
mgr, ax, mgr.axes[ax][[0, 1, 2, 0, 1, 2]], fill_value
|
|
)
|
|
|
|
@pytest.mark.parametrize("mgr", MANAGERS)
|
|
@pytest.mark.parametrize("fill_value", [None, np.nan, 100.0])
|
|
def test_reindex_indexer(self, fill_value, mgr):
|
|
def assert_reindex_indexer_is_ok(mgr, axis, new_labels, indexer, fill_value):
|
|
mat = mgr.as_array()
|
|
reindexed_mat = algos.take_nd(mat, indexer, axis, fill_value=fill_value)
|
|
reindexed = mgr.reindex_indexer(
|
|
new_labels, indexer, axis, fill_value=fill_value
|
|
)
|
|
tm.assert_numpy_array_equal(
|
|
reindexed_mat, reindexed.as_array(), check_dtype=False
|
|
)
|
|
tm.assert_index_equal(reindexed.axes[axis], new_labels)
|
|
|
|
for ax in range(mgr.ndim):
|
|
assert_reindex_indexer_is_ok(mgr, ax, pd.Index([]), [], fill_value)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr, ax, mgr.axes[ax], np.arange(mgr.shape[ax]), fill_value
|
|
)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr,
|
|
ax,
|
|
pd.Index(["foo"] * mgr.shape[ax]),
|
|
np.arange(mgr.shape[ax]),
|
|
fill_value,
|
|
)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr, ax, mgr.axes[ax][::-1], np.arange(mgr.shape[ax]), fill_value,
|
|
)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr, ax, mgr.axes[ax], np.arange(mgr.shape[ax])[::-1], fill_value,
|
|
)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr, ax, pd.Index(["foo", "bar", "baz"]), [0, 0, 0], fill_value
|
|
)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr, ax, pd.Index(["foo", "bar", "baz"]), [-1, 0, -1], fill_value,
|
|
)
|
|
assert_reindex_indexer_is_ok(
|
|
mgr,
|
|
ax,
|
|
pd.Index(["foo", mgr.axes[ax][0], "baz"]),
|
|
[-1, -1, -1],
|
|
fill_value,
|
|
)
|
|
|
|
if mgr.shape[ax] >= 3:
|
|
assert_reindex_indexer_is_ok(
|
|
mgr, ax, pd.Index(["foo", "bar", "baz"]), [0, 1, 2], fill_value,
|
|
)
|
|
|
|
|
|
class TestBlockPlacement:
|
|
@pytest.mark.parametrize(
|
|
"slc, expected",
|
|
[
|
|
(slice(0, 4), 4),
|
|
(slice(0, 4, 2), 2),
|
|
(slice(0, 3, 2), 2),
|
|
(slice(0, 1, 2), 1),
|
|
(slice(1, 0, -1), 1),
|
|
],
|
|
)
|
|
def test_slice_len(self, slc, expected):
|
|
assert len(BlockPlacement(slc)) == expected
|
|
|
|
@pytest.mark.parametrize("slc", [slice(1, 1, 0), slice(1, 2, 0)])
|
|
def test_zero_step_raises(self, slc):
|
|
msg = "slice step cannot be zero"
|
|
with pytest.raises(ValueError, match=msg):
|
|
BlockPlacement(slc)
|
|
|
|
@pytest.mark.parametrize(
|
|
"slc",
|
|
[
|
|
slice(None, None),
|
|
slice(10, None),
|
|
slice(None, None, -1),
|
|
slice(None, 10, -1),
|
|
# These are "unbounded" because negative index will
|
|
# change depending on container shape.
|
|
slice(-1, None),
|
|
slice(None, -1),
|
|
slice(-1, -1),
|
|
slice(-1, None, -1),
|
|
slice(None, -1, -1),
|
|
slice(-1, -1, -1),
|
|
],
|
|
)
|
|
def test_unbounded_slice_raises(self, slc):
|
|
msg = "unbounded slice"
|
|
with pytest.raises(ValueError, match=msg):
|
|
BlockPlacement(slc)
|
|
|
|
@pytest.mark.parametrize(
|
|
"slc",
|
|
[
|
|
slice(0, 0),
|
|
slice(100, 0),
|
|
slice(100, 100),
|
|
slice(100, 100, -1),
|
|
slice(0, 100, -1),
|
|
],
|
|
)
|
|
def test_not_slice_like_slices(self, slc):
|
|
assert not BlockPlacement(slc).is_slice_like
|
|
|
|
@pytest.mark.parametrize(
|
|
"arr, slc",
|
|
[
|
|
([0], slice(0, 1, 1)),
|
|
([100], slice(100, 101, 1)),
|
|
([0, 1, 2], slice(0, 3, 1)),
|
|
([0, 5, 10], slice(0, 15, 5)),
|
|
([0, 100], slice(0, 200, 100)),
|
|
([2, 1], slice(2, 0, -1)),
|
|
],
|
|
)
|
|
def test_array_to_slice_conversion(self, arr, slc):
|
|
assert BlockPlacement(arr).as_slice == slc
|
|
|
|
@pytest.mark.parametrize(
|
|
"arr",
|
|
[
|
|
[],
|
|
[-1],
|
|
[-1, -2, -3],
|
|
[-10],
|
|
[-1],
|
|
[-1, 0, 1, 2],
|
|
[-2, 0, 2, 4],
|
|
[1, 0, -1],
|
|
[1, 1, 1],
|
|
],
|
|
)
|
|
def test_not_slice_like_arrays(self, arr):
|
|
assert not BlockPlacement(arr).is_slice_like
|
|
|
|
@pytest.mark.parametrize(
|
|
"slc, expected",
|
|
[(slice(0, 3), [0, 1, 2]), (slice(0, 0), []), (slice(3, 0), [])],
|
|
)
|
|
def test_slice_iter(self, slc, expected):
|
|
assert list(BlockPlacement(slc)) == expected
|
|
|
|
@pytest.mark.parametrize(
|
|
"slc, arr",
|
|
[
|
|
(slice(0, 3), [0, 1, 2]),
|
|
(slice(0, 0), []),
|
|
(slice(3, 0), []),
|
|
(slice(3, 0, -1), [3, 2, 1]),
|
|
],
|
|
)
|
|
def test_slice_to_array_conversion(self, slc, arr):
|
|
tm.assert_numpy_array_equal(
|
|
BlockPlacement(slc).as_array, np.asarray(arr, dtype=np.int64)
|
|
)
|
|
|
|
def test_blockplacement_add(self):
|
|
bpl = BlockPlacement(slice(0, 5))
|
|
assert bpl.add(1).as_slice == slice(1, 6, 1)
|
|
assert bpl.add(np.arange(5)).as_slice == slice(0, 10, 2)
|
|
assert list(bpl.add(np.arange(5, 0, -1))) == [5, 5, 5, 5, 5]
|
|
|
|
@pytest.mark.parametrize(
|
|
"val, inc, expected",
|
|
[
|
|
(slice(0, 0), 0, []),
|
|
(slice(1, 4), 0, [1, 2, 3]),
|
|
(slice(3, 0, -1), 0, [3, 2, 1]),
|
|
([1, 2, 4], 0, [1, 2, 4]),
|
|
(slice(0, 0), 10, []),
|
|
(slice(1, 4), 10, [11, 12, 13]),
|
|
(slice(3, 0, -1), 10, [13, 12, 11]),
|
|
([1, 2, 4], 10, [11, 12, 14]),
|
|
(slice(0, 0), -1, []),
|
|
(slice(1, 4), -1, [0, 1, 2]),
|
|
([1, 2, 4], -1, [0, 1, 3]),
|
|
],
|
|
)
|
|
def test_blockplacement_add_int(self, val, inc, expected):
|
|
assert list(BlockPlacement(val).add(inc)) == expected
|
|
|
|
@pytest.mark.parametrize("val", [slice(1, 4), [1, 2, 4]])
|
|
def test_blockplacement_add_int_raises(self, val):
|
|
msg = "iadd causes length change"
|
|
with pytest.raises(ValueError, match=msg):
|
|
BlockPlacement(val).add(-10)
|
|
|
|
|
|
class DummyElement:
|
|
def __init__(self, value, dtype):
|
|
self.value = value
|
|
self.dtype = np.dtype(dtype)
|
|
|
|
def __array__(self):
|
|
return np.array(self.value, dtype=self.dtype)
|
|
|
|
def __str__(self) -> str:
|
|
return f"DummyElement({self.value}, {self.dtype})"
|
|
|
|
def __repr__(self) -> str:
|
|
return str(self)
|
|
|
|
def astype(self, dtype, copy=False):
|
|
self.dtype = dtype
|
|
return self
|
|
|
|
def view(self, dtype):
|
|
return type(self)(self.value.view(dtype), dtype)
|
|
|
|
def any(self, axis=None):
|
|
return bool(self.value)
|
|
|
|
|
|
class TestCanHoldElement:
|
|
def test_datetime_block_can_hold_element(self):
|
|
block = create_block("datetime", [0])
|
|
|
|
# We will check that block._can_hold_element iff arr.__setitem__ works
|
|
arr = pd.array(block.values.ravel())
|
|
|
|
# coerce None
|
|
assert block._can_hold_element(None)
|
|
arr[0] = None
|
|
assert arr[0] is pd.NaT
|
|
|
|
# coerce different types of datetime objects
|
|
vals = [np.datetime64("2010-10-10"), datetime(2010, 10, 10)]
|
|
for val in vals:
|
|
assert block._can_hold_element(val)
|
|
arr[0] = val
|
|
|
|
val = date(2010, 10, 10)
|
|
assert not block._can_hold_element(val)
|
|
|
|
msg = (
|
|
"'value' should be a 'Timestamp', 'NaT', "
|
|
"or array of those. Got 'date' instead."
|
|
)
|
|
with pytest.raises(TypeError, match=msg):
|
|
arr[0] = val
|
|
|
|
@pytest.mark.parametrize(
|
|
"value, dtype",
|
|
[
|
|
(1, "i8"),
|
|
(1.0, "f8"),
|
|
(2 ** 63, "f8"),
|
|
(1j, "complex128"),
|
|
(2 ** 63, "complex128"),
|
|
(True, "bool"),
|
|
(np.timedelta64(20, "ns"), "<m8[ns]"),
|
|
(np.datetime64(20, "ns"), "<M8[ns]"),
|
|
],
|
|
)
|
|
@pytest.mark.parametrize(
|
|
"op",
|
|
[
|
|
operator.add,
|
|
operator.sub,
|
|
operator.mul,
|
|
operator.truediv,
|
|
operator.mod,
|
|
operator.pow,
|
|
],
|
|
ids=lambda x: x.__name__,
|
|
)
|
|
def test_binop_other(self, op, value, dtype):
|
|
skip = {
|
|
(operator.add, "bool"),
|
|
(operator.sub, "bool"),
|
|
(operator.mul, "bool"),
|
|
(operator.truediv, "bool"),
|
|
(operator.mod, "i8"),
|
|
(operator.mod, "complex128"),
|
|
(operator.pow, "bool"),
|
|
}
|
|
if (op, dtype) in skip:
|
|
pytest.skip(f"Invalid combination {op},{dtype}")
|
|
|
|
e = DummyElement(value, dtype)
|
|
s = pd.DataFrame({"A": [e.value, e.value]}, dtype=e.dtype)
|
|
|
|
invalid = {
|
|
(operator.pow, "<M8[ns]"),
|
|
(operator.mod, "<M8[ns]"),
|
|
(operator.truediv, "<M8[ns]"),
|
|
(operator.mul, "<M8[ns]"),
|
|
(operator.add, "<M8[ns]"),
|
|
(operator.pow, "<m8[ns]"),
|
|
(operator.mul, "<m8[ns]"),
|
|
}
|
|
|
|
if (op, dtype) in invalid:
|
|
msg = (
|
|
None
|
|
if (dtype == "<M8[ns]" and op == operator.add)
|
|
or (dtype == "<m8[ns]" and op == operator.mul)
|
|
else (
|
|
f"cannot perform __{op.__name__}__ with this "
|
|
"index type: (DatetimeArray|TimedeltaArray)"
|
|
)
|
|
)
|
|
|
|
with pytest.raises(TypeError, match=msg):
|
|
op(s, e.value)
|
|
else:
|
|
# FIXME: Since dispatching to Series, this test no longer
|
|
# asserts anything meaningful
|
|
result = op(s, e.value).dtypes
|
|
expected = op(s, value).dtypes
|
|
tm.assert_series_equal(result, expected)
|
|
|
|
|
|
class TestShouldStore:
|
|
def test_should_store_categorical(self):
|
|
cat = pd.Categorical(["A", "B", "C"])
|
|
df = pd.DataFrame(cat)
|
|
blk = df._mgr.blocks[0]
|
|
|
|
# matching dtype
|
|
assert blk.should_store(cat)
|
|
assert blk.should_store(cat[:-1])
|
|
|
|
# different dtype
|
|
assert not blk.should_store(cat.as_ordered())
|
|
|
|
# ndarray instead of Categorical
|
|
assert not blk.should_store(np.asarray(cat))
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"typestr, holder",
|
|
[
|
|
("category", Categorical),
|
|
("M8[ns]", DatetimeArray),
|
|
("M8[ns, US/Central]", DatetimeArray),
|
|
("m8[ns]", TimedeltaArray),
|
|
("sparse", SparseArray),
|
|
],
|
|
)
|
|
def test_holder(typestr, holder):
|
|
blk = create_block(typestr, [1])
|
|
assert blk._holder is holder
|
|
|
|
|
|
def test_validate_ndim():
|
|
values = np.array([1.0, 2.0])
|
|
placement = slice(2)
|
|
msg = r"Wrong number of dimensions. values.ndim != ndim \[1 != 2\]"
|
|
|
|
with pytest.raises(ValueError, match=msg):
|
|
make_block(values, placement, ndim=2)
|
|
|
|
|
|
def test_block_shape():
|
|
idx = pd.Index([0, 1, 2, 3, 4])
|
|
a = pd.Series([1, 2, 3]).reindex(idx)
|
|
b = pd.Series(pd.Categorical([1, 2, 3])).reindex(idx)
|
|
|
|
assert a._mgr.blocks[0].mgr_locs.indexer == b._mgr.blocks[0].mgr_locs.indexer
|
|
|
|
|
|
def test_make_block_no_pandas_array():
|
|
# https://github.com/pandas-dev/pandas/pull/24866
|
|
arr = pd.arrays.PandasArray(np.array([1, 2]))
|
|
|
|
# PandasArray, no dtype
|
|
result = make_block(arr, slice(len(arr)))
|
|
assert result.is_integer is True
|
|
assert result.is_extension is False
|
|
|
|
# PandasArray, PandasDtype
|
|
result = make_block(arr, slice(len(arr)), dtype=arr.dtype)
|
|
assert result.is_integer is True
|
|
assert result.is_extension is False
|
|
|
|
# ndarray, PandasDtype
|
|
result = make_block(arr.to_numpy(), slice(len(arr)), dtype=arr.dtype)
|
|
assert result.is_integer is True
|
|
assert result.is_extension is False
|
|
|
|
|
|
def test_dataframe_not_equal():
|
|
# see GH28839
|
|
df1 = pd.DataFrame({"a": [1, 2], "b": ["s", "d"]})
|
|
df2 = pd.DataFrame({"a": ["s", "d"], "b": [1, 2]})
|
|
assert df1.equals(df2) is False
|
|
|
|
|
|
def test_missing_unicode_key():
|
|
df = DataFrame({"a": [1]})
|
|
with pytest.raises(KeyError, match="\u05d0"):
|
|
df.loc[:, "\u05d0"] # should not raise UnicodeEncodeError
|
|
|
|
|
|
def test_set_change_dtype_slice():
|
|
# GH#8850
|
|
cols = MultiIndex.from_tuples([("1st", "a"), ("2nd", "b"), ("3rd", "c")])
|
|
df = DataFrame([[1.0, 2, 3], [4.0, 5, 6]], columns=cols)
|
|
df["2nd"] = df["2nd"] * 2.0
|
|
|
|
blocks = df._to_dict_of_blocks()
|
|
assert sorted(blocks.keys()) == ["float64", "int64"]
|
|
tm.assert_frame_equal(
|
|
blocks["float64"], DataFrame([[1.0, 4.0], [4.0, 10.0]], columns=cols[:2])
|
|
)
|
|
tm.assert_frame_equal(blocks["int64"], DataFrame([[3], [6]], columns=cols[2:]))
|
|
|
|
|
|
def test_interleave_non_unique_cols():
|
|
df = DataFrame(
|
|
[[pd.Timestamp("20130101"), 3.5], [pd.Timestamp("20130102"), 4.5]],
|
|
columns=["x", "x"],
|
|
index=[1, 2],
|
|
)
|
|
|
|
df_unique = df.copy()
|
|
df_unique.columns = ["x", "y"]
|
|
assert df_unique.values.shape == df.values.shape
|
|
tm.assert_numpy_array_equal(df_unique.values[0], df.values[0])
|
|
tm.assert_numpy_array_equal(df_unique.values[1], df.values[1])
|
|
|
|
|
|
def test_single_block_manager_fastpath_deprecated():
|
|
# GH#33092
|
|
ser = pd.Series(range(3))
|
|
blk = ser._data.blocks[0]
|
|
with tm.assert_produces_warning(FutureWarning):
|
|
SingleBlockManager(blk, ser.index, fastpath=True)
|
|
|