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
|
# BSD 3-Clause License; see https://github.com/scikit-hep/awkward/blob/main/LICENSE
# ruff: noqa: E402
from __future__ import annotations
import os
import numpy as np
import pytest
pa = pytest.importorskip("pyarrow")
pq = pytest.importorskip("pyarrow.parquet")
import awkward as ak
from awkward._connect.pyarrow.table_conv import (
AWKWARD_INFO_KEY,
array_with_replacement_type,
awkward_arrow_field_to_native,
collect_ak_arr_type_metadata,
convert_awkward_arrow_table_to_native,
convert_native_arrow_table_to_awkward,
native_arrow_field_to_akarraytype,
)
from awkward.operations import to_list
nested_ints = ak.Array([[[[[1, 2, 3], [], [4, 5]] * 5] * 3] * 2])
struct_array = ak.Array(
[
{"lbl": "item 1", "idx": 11, "ids": [1, 2, 3]},
{"lbl": "item 2", "idx": 12, "ids": [51, 52]},
{"lbl": "item 3", "idx": 13, "ids": [61, 62, 63, 64]},
],
with_name="my_items",
)
with_nulls = ak.Array([[11, None], [23], None])
option_types = ak.Array(
[[1, 2], [[10, 11, None], None, [3, 4, 5], ["one", "two"]], None]
)
(indexed, indexed_tuple) = (
# Unique for this test, contains nested types within the
ak.contents.IndexedArray(
ak.index.Index64(np.array([2, 1, 0], dtype=np.uint64)),
ak.contents.RecordArray(
[
ak.contents.ByteMaskedArray(
ak.index.Index8(np.array([False, True, False]).view(np.int8)),
ak.contents.NumpyArray(np.array([1.1, 2.2, 3.3])),
valid_when=False,
),
ak.contents.UnmaskedArray(
ak.contents.ListOffsetArray(
ak.index.Index32(np.array([0, 3, 3, 5], dtype=np.int32)),
ak.contents.NumpyArray(
np.array([1.1, 2.2, 3.3, 4.4, 5.5]),
),
),
),
],
None if is_tuple else ["x", "y"],
),
parameters={"__array__": "categorical", "foo": "thap"},
)
for is_tuple in (False, True)
)
(records, records_tuple) = (
ak.contents.RecordArray(
[
ak.contents.NumpyArray(
np.array([1.1, 2.2, 3.3]), parameters={"foo": "inner1"}
),
ak.contents.ListOffsetArray(
ak.index.Index32(np.array([0, 3, 3, 5], dtype=np.int32)),
ak.contents.NumpyArray(
np.array([1.1, 2.2, 3.3, 4.4, 5.5]), parameters={"foo": "inner2"}
),
),
],
None if is_tuple else ["x", "y"],
parameters={"foo": "outer"},
)
for is_tuple in (False, True)
)
@pytest.mark.parametrize(
"akarray, as_dict",
[
(nested_ints, False),
(struct_array, False),
(with_nulls, False),
(option_types, False),
(indexed, True),
(indexed_tuple, True),
(records, True),
(records_tuple, True),
(indexed, False),
(indexed_tuple, False),
(records, False),
(records_tuple, False),
],
)
def test_array_conversions(akarray, as_dict):
arrow_natv = ak.to_arrow(
akarray, categorical_as_dictionary=as_dict, extensionarray=False
)
arrow_extn = ak.to_arrow(
akarray, categorical_as_dictionary=as_dict, extensionarray=True
)
extn_field = pa.field("test_field", arrow_extn.type)
# Convert to native pyarrow types:
metadata = collect_ak_arr_type_metadata(extn_field)
conv_natv_field = awkward_arrow_field_to_native(extn_field)
assert conv_natv_field.type == arrow_natv.type
as_natv = array_with_replacement_type(arrow_extn, conv_natv_field.type)
assert as_natv == arrow_natv
# Complete a round-trip, back to AwkwardArrowArray
conv_extn_field = native_arrow_field_to_akarraytype(conv_natv_field, metadata)
assert conv_extn_field.type == arrow_extn.type
assert (
conv_extn_field.type._metadata_as_dict() == arrow_extn.type._metadata_as_dict()
)
conv_metadata = collect_ak_arr_type_metadata(conv_extn_field)
assert conv_metadata == metadata
# assert_equal_arrowextntypes(conv_extn_field.type, arrow_extn.type)
as_extn = array_with_replacement_type(as_natv, conv_extn_field.type)
assert as_extn == arrow_extn
assert as_extn.type._metadata_as_dict() == arrow_extn.type._metadata_as_dict()
# And back to Awkward array
rt_array = ak.from_arrow(as_extn, highlevel=True)
assert to_list(rt_array) == to_list(akarray)
def test_table_conversion():
ak_tbl_like = ak.Array(
{
"struct_array": struct_array,
"with_nulls": with_nulls,
"option_types": option_types,
"indexed": indexed,
}
)
natv_tbl = ak.to_arrow_table(
ak_tbl_like, categorical_as_dictionary=True, extensionarray=False
)
extn_tbl = ak.to_arrow_table(
ak_tbl_like, categorical_as_dictionary=True, extensionarray=True
)
conv_natv_tbl = convert_awkward_arrow_table_to_native(extn_tbl)
assert (
conv_natv_tbl.schema == natv_tbl.schema
) # This comparison does not include metadata
assert conv_natv_tbl == natv_tbl
assert AWKWARD_INFO_KEY in conv_natv_tbl.schema.metadata
assert len(conv_natv_tbl["struct_array"].chunks[0].field(2).buffers()) == 4
# This verifies that our conversions are zero-copy for the table data
assert (
extn_tbl["struct_array"].chunks[0].storage.field(2).buffers()[3].address
== conv_natv_tbl["struct_array"].chunks[0].field(2).buffers()[3].address
)
conv_extn_tbl = convert_native_arrow_table_to_awkward(conv_natv_tbl)
assert conv_extn_tbl.schema == extn_tbl.schema
assert conv_extn_tbl == extn_tbl # Full Round-trip verification
assert AWKWARD_INFO_KEY not in conv_extn_tbl.schema.metadata
assert (
extn_tbl["struct_array"].chunks[0].storage.field(2).buffers()[3].address
== conv_extn_tbl["struct_array"].chunks[0].storage.field(2).buffers()[3].address
)
def test_selective_parquet(tmp_path):
filename = os.path.join(tmp_path, "whatever.parquet")
ak_tbl = ak.Array(
{
"with_nulls": with_nulls,
"struct_array": struct_array,
"indexed": indexed,
}
)
ak.to_parquet(ak_tbl, filename)
tbl_tr = ak.from_parquet(filename, columns=["struct_array", "indexed"])
assert to_list(tbl_tr["struct_array"]) == to_list(ak_tbl["struct_array"])
|
