1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
|
from __future__ import annotations
import operator
from functools import reduce
from itertools import chain
from typing import TYPE_CHECKING, Literal, Sequence
import pyarrow as pa
import pyarrow.compute as pc
from narwhals._arrow.dataframe import ArrowDataFrame
from narwhals._arrow.expr import ArrowExpr
from narwhals._arrow.selectors import ArrowSelectorNamespace
from narwhals._arrow.series import ArrowSeries
from narwhals._arrow.utils import (
align_series_full_broadcast,
cast_to_comparable_string_types,
)
from narwhals._compliant import CompliantThen, EagerNamespace, EagerWhen
from narwhals._expression_parsing import (
combine_alias_output_names,
combine_evaluate_output_names,
)
from narwhals._utils import Implementation
if TYPE_CHECKING:
from narwhals._arrow.typing import Incomplete
from narwhals._utils import Version
from narwhals.typing import IntoDType, NonNestedLiteral
class ArrowNamespace(EagerNamespace[ArrowDataFrame, ArrowSeries, ArrowExpr, pa.Table]):
@property
def _dataframe(self) -> type[ArrowDataFrame]:
return ArrowDataFrame
@property
def _expr(self) -> type[ArrowExpr]:
return ArrowExpr
@property
def _series(self) -> type[ArrowSeries]:
return ArrowSeries
# --- not in spec ---
def __init__(self, *, backend_version: tuple[int, ...], version: Version) -> None:
self._backend_version = backend_version
self._implementation = Implementation.PYARROW
self._version = version
def len(self) -> ArrowExpr:
# coverage bug? this is definitely hit
return self._expr( # pragma: no cover
lambda df: [
ArrowSeries.from_iterable([len(df.native)], name="len", context=self)
],
depth=0,
function_name="len",
evaluate_output_names=lambda _df: ["len"],
alias_output_names=None,
backend_version=self._backend_version,
version=self._version,
)
def lit(self, value: NonNestedLiteral, dtype: IntoDType | None) -> ArrowExpr:
def _lit_arrow_series(_: ArrowDataFrame) -> ArrowSeries:
arrow_series = ArrowSeries.from_iterable(
data=[value], name="literal", context=self
)
if dtype:
return arrow_series.cast(dtype)
return arrow_series
return self._expr(
lambda df: [_lit_arrow_series(df)],
depth=0,
function_name="lit",
evaluate_output_names=lambda _df: ["literal"],
alias_output_names=None,
backend_version=self._backend_version,
version=self._version,
)
def all_horizontal(self, *exprs: ArrowExpr) -> ArrowExpr:
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
series = chain.from_iterable(expr(df) for expr in exprs)
return [reduce(operator.and_, align_series_full_broadcast(*series))]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="all_horizontal",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
def any_horizontal(self, *exprs: ArrowExpr) -> ArrowExpr:
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
series = chain.from_iterable(expr(df) for expr in exprs)
return [reduce(operator.or_, align_series_full_broadcast(*series))]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="any_horizontal",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
def sum_horizontal(self, *exprs: ArrowExpr) -> ArrowExpr:
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
it = chain.from_iterable(expr(df) for expr in exprs)
series = (s.fill_null(0, strategy=None, limit=None) for s in it)
return [reduce(operator.add, align_series_full_broadcast(*series))]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="sum_horizontal",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
def mean_horizontal(self, *exprs: ArrowExpr) -> ArrowExpr:
int_64 = self._version.dtypes.Int64()
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
expr_results = list(chain.from_iterable(expr(df) for expr in exprs))
series = align_series_full_broadcast(
*(s.fill_null(0, strategy=None, limit=None) for s in expr_results)
)
non_na = align_series_full_broadcast(
*(1 - s.is_null().cast(int_64) for s in expr_results)
)
return [reduce(operator.add, series) / reduce(operator.add, non_na)]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="mean_horizontal",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
def min_horizontal(self, *exprs: ArrowExpr) -> ArrowExpr:
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
init_series, *series = list(chain.from_iterable(expr(df) for expr in exprs))
init_series, *series = align_series_full_broadcast(init_series, *series)
native_series = reduce(
pc.min_element_wise, [s.native for s in series], init_series.native
)
return [
ArrowSeries(
native_series,
name=init_series.name,
backend_version=self._backend_version,
version=self._version,
)
]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="min_horizontal",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
def max_horizontal(self, *exprs: ArrowExpr) -> ArrowExpr:
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
init_series, *series = list(chain.from_iterable(expr(df) for expr in exprs))
init_series, *series = align_series_full_broadcast(init_series, *series)
native_series = reduce(
pc.max_element_wise, [s.native for s in series], init_series.native
)
return [
ArrowSeries(
native_series,
name=init_series.name,
backend_version=self._backend_version,
version=self._version,
)
]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="max_horizontal",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
def _concat_diagonal(self, dfs: Sequence[pa.Table], /) -> pa.Table:
if self._backend_version >= (14,):
return pa.concat_tables(dfs, promote_options="default")
return pa.concat_tables(dfs, promote=True) # pragma: no cover
def _concat_horizontal(self, dfs: Sequence[pa.Table], /) -> pa.Table:
names = list(chain.from_iterable(df.column_names for df in dfs))
arrays = list(chain.from_iterable(df.itercolumns() for df in dfs))
return pa.Table.from_arrays(arrays, names=names)
def _concat_vertical(self, dfs: Sequence[pa.Table], /) -> pa.Table:
cols_0 = dfs[0].column_names
for i, df in enumerate(dfs[1:], start=1):
cols_current = df.column_names
if cols_current != cols_0:
msg = (
"unable to vstack, column names don't match:\n"
f" - dataframe 0: {cols_0}\n"
f" - dataframe {i}: {cols_current}\n"
)
raise TypeError(msg)
return pa.concat_tables(dfs)
@property
def selectors(self) -> ArrowSelectorNamespace:
return ArrowSelectorNamespace.from_namespace(self)
def when(self, predicate: ArrowExpr) -> ArrowWhen:
return ArrowWhen.from_expr(predicate, context=self)
def concat_str(
self, *exprs: ArrowExpr, separator: str, ignore_nulls: bool
) -> ArrowExpr:
def func(df: ArrowDataFrame) -> list[ArrowSeries]:
compliant_series_list = align_series_full_broadcast(
*(chain.from_iterable(expr(df) for expr in exprs))
)
name = compliant_series_list[0].name
null_handling: Literal["skip", "emit_null"] = (
"skip" if ignore_nulls else "emit_null"
)
it, separator_scalar = cast_to_comparable_string_types(
*(s.native for s in compliant_series_list), separator=separator
)
# NOTE: stubs indicate `separator` must also be a `ChunkedArray`
# Reality: `str` is fine
concat_str: Incomplete = pc.binary_join_element_wise
compliant = self._series(
concat_str(*it, separator_scalar, null_handling=null_handling),
name=name,
backend_version=self._backend_version,
version=self._version,
)
return [compliant]
return self._expr._from_callable(
func=func,
depth=max(x._depth for x in exprs) + 1,
function_name="concat_str",
evaluate_output_names=combine_evaluate_output_names(*exprs),
alias_output_names=combine_alias_output_names(*exprs),
context=self,
)
class ArrowWhen(EagerWhen[ArrowDataFrame, ArrowSeries, ArrowExpr]):
@property
def _then(self) -> type[ArrowThen]:
return ArrowThen
def _if_then_else(
self, when: ArrowSeries, then: ArrowSeries, otherwise: ArrowSeries | None, /
) -> ArrowSeries:
if otherwise is None:
when, then = align_series_full_broadcast(when, then)
res_native = pc.if_else(
when.native, then.native, pa.nulls(len(when.native), then.native.type)
)
else:
when, then, otherwise = align_series_full_broadcast(when, then, otherwise)
res_native = pc.if_else(when.native, then.native, otherwise.native)
return then._with_native(res_native)
class ArrowThen(CompliantThen[ArrowDataFrame, ArrowSeries, ArrowExpr], ArrowExpr): ...
|
