polars-df 0.10.0-x86_64-linux-musl → 0.11.0-x86_64-linux-musl
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +11 -0
- data/Cargo.lock +90 -48
- data/LICENSE-THIRD-PARTY.txt +152 -79
- data/README.md +6 -6
- data/lib/polars/3.1/polars.so +0 -0
- data/lib/polars/3.2/polars.so +0 -0
- data/lib/polars/3.3/polars.so +0 -0
- data/lib/polars/batched_csv_reader.rb +9 -3
- data/lib/polars/convert.rb +6 -1
- data/lib/polars/data_frame.rb +83 -302
- data/lib/polars/date_time_expr.rb +1 -0
- data/lib/polars/date_time_name_space.rb +5 -1
- data/lib/polars/dynamic_group_by.rb +2 -2
- data/lib/polars/exceptions.rb +4 -0
- data/lib/polars/expr.rb +1134 -20
- data/lib/polars/functions/range/date_range.rb +92 -0
- data/lib/polars/functions/range/datetime_range.rb +149 -0
- data/lib/polars/functions/range/time_range.rb +141 -0
- data/lib/polars/group_by.rb +88 -23
- data/lib/polars/io/avro.rb +24 -0
- data/lib/polars/{io.rb → io/csv.rb} +296 -490
- data/lib/polars/io/database.rb +73 -0
- data/lib/polars/io/ipc.rb +247 -0
- data/lib/polars/io/json.rb +18 -0
- data/lib/polars/io/ndjson.rb +69 -0
- data/lib/polars/io/parquet.rb +226 -0
- data/lib/polars/lazy_frame.rb +23 -166
- data/lib/polars/lazy_group_by.rb +100 -3
- data/lib/polars/rolling_group_by.rb +2 -2
- data/lib/polars/series.rb +2 -2
- data/lib/polars/string_expr.rb +37 -36
- data/lib/polars/utils.rb +35 -1
- data/lib/polars/version.rb +1 -1
- data/lib/polars.rb +9 -1
- metadata +12 -4
data/lib/polars/lazy_frame.rb
CHANGED
@@ -27,149 +27,6 @@ module Polars
|
|
27
27
|
ldf
|
28
28
|
end
|
29
29
|
|
30
|
-
# @private
|
31
|
-
def self._scan_csv(
|
32
|
-
file,
|
33
|
-
has_header: true,
|
34
|
-
sep: ",",
|
35
|
-
comment_char: nil,
|
36
|
-
quote_char: '"',
|
37
|
-
skip_rows: 0,
|
38
|
-
dtypes: nil,
|
39
|
-
null_values: nil,
|
40
|
-
ignore_errors: false,
|
41
|
-
cache: true,
|
42
|
-
with_column_names: nil,
|
43
|
-
infer_schema_length: 100,
|
44
|
-
n_rows: nil,
|
45
|
-
encoding: "utf8",
|
46
|
-
low_memory: false,
|
47
|
-
rechunk: true,
|
48
|
-
skip_rows_after_header: 0,
|
49
|
-
row_count_name: nil,
|
50
|
-
row_count_offset: 0,
|
51
|
-
parse_dates: false,
|
52
|
-
eol_char: "\n",
|
53
|
-
truncate_ragged_lines: true
|
54
|
-
)
|
55
|
-
dtype_list = nil
|
56
|
-
if !dtypes.nil?
|
57
|
-
dtype_list = []
|
58
|
-
dtypes.each do |k, v|
|
59
|
-
dtype_list << [k, Utils.rb_type_to_dtype(v)]
|
60
|
-
end
|
61
|
-
end
|
62
|
-
processed_null_values = Utils._process_null_values(null_values)
|
63
|
-
|
64
|
-
_from_rbldf(
|
65
|
-
RbLazyFrame.new_from_csv(
|
66
|
-
file,
|
67
|
-
sep,
|
68
|
-
has_header,
|
69
|
-
ignore_errors,
|
70
|
-
skip_rows,
|
71
|
-
n_rows,
|
72
|
-
cache,
|
73
|
-
dtype_list,
|
74
|
-
low_memory,
|
75
|
-
comment_char,
|
76
|
-
quote_char,
|
77
|
-
processed_null_values,
|
78
|
-
infer_schema_length,
|
79
|
-
with_column_names,
|
80
|
-
rechunk,
|
81
|
-
skip_rows_after_header,
|
82
|
-
encoding,
|
83
|
-
Utils._prepare_row_count_args(row_count_name, row_count_offset),
|
84
|
-
parse_dates,
|
85
|
-
eol_char,
|
86
|
-
truncate_ragged_lines
|
87
|
-
)
|
88
|
-
)
|
89
|
-
end
|
90
|
-
|
91
|
-
# @private
|
92
|
-
def self._scan_parquet(
|
93
|
-
file,
|
94
|
-
n_rows: nil,
|
95
|
-
cache: true,
|
96
|
-
parallel: "auto",
|
97
|
-
rechunk: true,
|
98
|
-
row_count_name: nil,
|
99
|
-
row_count_offset: 0,
|
100
|
-
storage_options: nil,
|
101
|
-
low_memory: false,
|
102
|
-
use_statistics: true,
|
103
|
-
hive_partitioning: true
|
104
|
-
)
|
105
|
-
_from_rbldf(
|
106
|
-
RbLazyFrame.new_from_parquet(
|
107
|
-
file,
|
108
|
-
[],
|
109
|
-
n_rows,
|
110
|
-
cache,
|
111
|
-
parallel,
|
112
|
-
rechunk,
|
113
|
-
Utils._prepare_row_count_args(row_count_name, row_count_offset),
|
114
|
-
low_memory,
|
115
|
-
use_statistics,
|
116
|
-
hive_partitioning,
|
117
|
-
nil
|
118
|
-
)
|
119
|
-
)
|
120
|
-
end
|
121
|
-
|
122
|
-
# @private
|
123
|
-
def self._scan_ipc(
|
124
|
-
file,
|
125
|
-
n_rows: nil,
|
126
|
-
cache: true,
|
127
|
-
rechunk: true,
|
128
|
-
row_count_name: nil,
|
129
|
-
row_count_offset: 0,
|
130
|
-
storage_options: nil,
|
131
|
-
memory_map: true
|
132
|
-
)
|
133
|
-
if Utils.pathlike?(file)
|
134
|
-
file = Utils.normalise_filepath(file)
|
135
|
-
end
|
136
|
-
|
137
|
-
_from_rbldf(
|
138
|
-
RbLazyFrame.new_from_ipc(
|
139
|
-
file,
|
140
|
-
n_rows,
|
141
|
-
cache,
|
142
|
-
rechunk,
|
143
|
-
Utils._prepare_row_count_args(row_count_name, row_count_offset),
|
144
|
-
memory_map
|
145
|
-
)
|
146
|
-
)
|
147
|
-
end
|
148
|
-
|
149
|
-
# @private
|
150
|
-
def self._scan_ndjson(
|
151
|
-
file,
|
152
|
-
infer_schema_length: nil,
|
153
|
-
batch_size: nil,
|
154
|
-
n_rows: nil,
|
155
|
-
low_memory: false,
|
156
|
-
rechunk: true,
|
157
|
-
row_count_name: nil,
|
158
|
-
row_count_offset: 0
|
159
|
-
)
|
160
|
-
_from_rbldf(
|
161
|
-
RbLazyFrame.new_from_ndjson(
|
162
|
-
file,
|
163
|
-
infer_schema_length,
|
164
|
-
batch_size,
|
165
|
-
n_rows,
|
166
|
-
low_memory,
|
167
|
-
rechunk,
|
168
|
-
Utils._prepare_row_count_args(row_count_name, row_count_offset)
|
169
|
-
)
|
170
|
-
)
|
171
|
-
end
|
172
|
-
|
173
30
|
# def self.from_json
|
174
31
|
# end
|
175
32
|
|
@@ -181,7 +38,7 @@ module Polars
|
|
181
38
|
# @return [LazyFrame]
|
182
39
|
def self.read_json(file)
|
183
40
|
if Utils.pathlike?(file)
|
184
|
-
file = Utils.
|
41
|
+
file = Utils.normalize_filepath(file)
|
185
42
|
end
|
186
43
|
|
187
44
|
Utils.wrap_ldf(RbLazyFrame.read_json(file))
|
@@ -288,7 +145,7 @@ module Polars
|
|
288
145
|
# @return [nil]
|
289
146
|
def write_json(file)
|
290
147
|
if Utils.pathlike?(file)
|
291
|
-
file = Utils.
|
148
|
+
file = Utils.normalize_filepath(file)
|
292
149
|
end
|
293
150
|
_ldf.write_json(file)
|
294
151
|
nil
|
@@ -1137,7 +994,7 @@ module Polars
|
|
1137
994
|
# # ┌─────────┐
|
1138
995
|
# # │ literal │
|
1139
996
|
# # │ --- │
|
1140
|
-
# # │
|
997
|
+
# # │ i32 │
|
1141
998
|
# # ╞═════════╡
|
1142
999
|
# # │ 0 │
|
1143
1000
|
# # │ 0 │
|
@@ -1449,16 +1306,16 @@ module Polars
|
|
1449
1306
|
# )
|
1450
1307
|
# # =>
|
1451
1308
|
# # shape: (4, 3)
|
1452
|
-
# #
|
1453
|
-
# # │ time ┆ time_count ┆ time_agg_list
|
1454
|
-
# # │ --- ┆ --- ┆ ---
|
1455
|
-
# # │ datetime[μs] ┆ u32 ┆ list[datetime[μs]]
|
1456
|
-
# #
|
1457
|
-
# # │ 2021-12-16 00:00:00 ┆ 2 ┆ [2021-12-16 00:00:00, 2021-12
|
1458
|
-
# # │ 2021-12-16 01:00:00 ┆ 2 ┆ [2021-12-16 01:00:00, 2021-12
|
1459
|
-
# # │ 2021-12-16 02:00:00 ┆ 2 ┆ [2021-12-16 02:00:00, 2021-12
|
1460
|
-
# # │ 2021-12-16 03:00:00 ┆ 1 ┆ [2021-12-16 03:00:00]
|
1461
|
-
# #
|
1309
|
+
# # ┌─────────────────────┬────────────┬─────────────────────────────────┐
|
1310
|
+
# # │ time ┆ time_count ┆ time_agg_list │
|
1311
|
+
# # │ --- ┆ --- ┆ --- │
|
1312
|
+
# # │ datetime[μs] ┆ u32 ┆ list[datetime[μs]] │
|
1313
|
+
# # ╞═════════════════════╪════════════╪═════════════════════════════════╡
|
1314
|
+
# # │ 2021-12-16 00:00:00 ┆ 2 ┆ [2021-12-16 00:00:00, 2021-12-… │
|
1315
|
+
# # │ 2021-12-16 01:00:00 ┆ 2 ┆ [2021-12-16 01:00:00, 2021-12-… │
|
1316
|
+
# # │ 2021-12-16 02:00:00 ┆ 2 ┆ [2021-12-16 02:00:00, 2021-12-… │
|
1317
|
+
# # │ 2021-12-16 03:00:00 ┆ 1 ┆ [2021-12-16 03:00:00] │
|
1318
|
+
# # └─────────────────────┴────────────┴─────────────────────────────────┘
|
1462
1319
|
#
|
1463
1320
|
# @example When closed="both" the time values at the window boundaries belong to 2 groups.
|
1464
1321
|
# df.group_by_dynamic("time", every: "1h", closed: "both").agg(
|
@@ -2236,16 +2093,16 @@ module Polars
|
|
2236
2093
|
# df.with_row_index.collect
|
2237
2094
|
# # =>
|
2238
2095
|
# # shape: (3, 3)
|
2239
|
-
# #
|
2240
|
-
# # │
|
2241
|
-
# # │ ---
|
2242
|
-
# # │ u32
|
2243
|
-
# #
|
2244
|
-
# # │ 0
|
2245
|
-
# # │ 1
|
2246
|
-
# # │ 2
|
2247
|
-
# #
|
2248
|
-
def with_row_index(name: "
|
2096
|
+
# # ┌───────┬─────┬─────┐
|
2097
|
+
# # │ index ┆ a ┆ b │
|
2098
|
+
# # │ --- ┆ --- ┆ --- │
|
2099
|
+
# # │ u32 ┆ i64 ┆ i64 │
|
2100
|
+
# # ╞═══════╪═════╪═════╡
|
2101
|
+
# # │ 0 ┆ 1 ┆ 2 │
|
2102
|
+
# # │ 1 ┆ 3 ┆ 4 │
|
2103
|
+
# # │ 2 ┆ 5 ┆ 6 │
|
2104
|
+
# # └───────┴─────┴─────┘
|
2105
|
+
def with_row_index(name: "index", offset: 0)
|
2249
2106
|
_from_rbldf(_ldf.with_row_index(name, offset))
|
2250
2107
|
end
|
2251
2108
|
alias_method :with_row_count, :with_row_index
|
data/lib/polars/lazy_group_by.rb
CHANGED
@@ -6,11 +6,108 @@ module Polars
|
|
6
6
|
@lgb = lgb
|
7
7
|
end
|
8
8
|
|
9
|
-
#
|
9
|
+
# Compute aggregations for each group of a group by operation.
|
10
|
+
#
|
11
|
+
# @param aggs [Array]
|
12
|
+
# Aggregations to compute for each group of the group by operation,
|
13
|
+
# specified as positional arguments.
|
14
|
+
# Accepts expression input. Strings are parsed as column names.
|
15
|
+
# @param named_aggs [Hash]
|
16
|
+
# Additional aggregations, specified as keyword arguments.
|
17
|
+
# The resulting columns will be renamed to the keyword used.
|
10
18
|
#
|
11
19
|
# @return [LazyFrame]
|
12
|
-
|
13
|
-
|
20
|
+
#
|
21
|
+
# @example Compute the aggregation of the columns for each group.
|
22
|
+
# ldf = Polars::DataFrame.new(
|
23
|
+
# {
|
24
|
+
# "a" => ["a", "b", "a", "b", "c"],
|
25
|
+
# "b" => [1, 2, 1, 3, 3],
|
26
|
+
# "c" => [5, 4, 3, 2, 1]
|
27
|
+
# }
|
28
|
+
# ).lazy
|
29
|
+
# ldf.group_by("a").agg(
|
30
|
+
# [Polars.col("b"), Polars.col("c")]
|
31
|
+
# ).collect
|
32
|
+
# # =>
|
33
|
+
# # shape: (3, 3)
|
34
|
+
# # ┌─────┬───────────┬───────────┐
|
35
|
+
# # │ a ┆ b ┆ c │
|
36
|
+
# # │ --- ┆ --- ┆ --- │
|
37
|
+
# # │ str ┆ list[i64] ┆ list[i64] │
|
38
|
+
# # ╞═════╪═══════════╪═══════════╡
|
39
|
+
# # │ a ┆ [1, 1] ┆ [5, 3] │
|
40
|
+
# # │ b ┆ [2, 3] ┆ [4, 2] │
|
41
|
+
# # │ c ┆ [3] ┆ [1] │
|
42
|
+
# # └─────┴───────────┴───────────┘
|
43
|
+
#
|
44
|
+
# @example Compute the sum of a column for each group.
|
45
|
+
# ldf.group_by("a").agg(
|
46
|
+
# Polars.col("b").sum
|
47
|
+
# ).collect
|
48
|
+
# # =>
|
49
|
+
# # shape: (3, 2)
|
50
|
+
# # ┌─────┬─────┐
|
51
|
+
# # │ a ┆ b │
|
52
|
+
# # │ --- ┆ --- │
|
53
|
+
# # │ str ┆ i64 │
|
54
|
+
# # ╞═════╪═════╡
|
55
|
+
# # │ a ┆ 2 │
|
56
|
+
# # │ b ┆ 5 │
|
57
|
+
# # │ c ┆ 3 │
|
58
|
+
# # └─────┴─────┘
|
59
|
+
#
|
60
|
+
# @example Compute multiple aggregates at once by passing a list of expressions.
|
61
|
+
# ldf.group_by("a").agg(
|
62
|
+
# [Polars.sum("b"), Polars.mean("c")]
|
63
|
+
# ).collect
|
64
|
+
# # =>
|
65
|
+
# # shape: (3, 3)
|
66
|
+
# # ┌─────┬─────┬─────┐
|
67
|
+
# # │ a ┆ b ┆ c │
|
68
|
+
# # │ --- ┆ --- ┆ --- │
|
69
|
+
# # │ str ┆ i64 ┆ f64 │
|
70
|
+
# # ╞═════╪═════╪═════╡
|
71
|
+
# # │ c ┆ 3 ┆ 1.0 │
|
72
|
+
# # │ a ┆ 2 ┆ 4.0 │
|
73
|
+
# # │ b ┆ 5 ┆ 3.0 │
|
74
|
+
# # └─────┴─────┴─────┘
|
75
|
+
#
|
76
|
+
# @example Or use positional arguments to compute multiple aggregations in the same way.
|
77
|
+
# ldf.group_by("a").agg(
|
78
|
+
# Polars.sum("b").name.suffix("_sum"),
|
79
|
+
# (Polars.col("c") ** 2).mean.name.suffix("_mean_squared")
|
80
|
+
# ).collect
|
81
|
+
# # =>
|
82
|
+
# # shape: (3, 3)
|
83
|
+
# # ┌─────┬───────┬────────────────┐
|
84
|
+
# # │ a ┆ b_sum ┆ c_mean_squared │
|
85
|
+
# # │ --- ┆ --- ┆ --- │
|
86
|
+
# # │ str ┆ i64 ┆ f64 │
|
87
|
+
# # ╞═════╪═══════╪════════════════╡
|
88
|
+
# # │ a ┆ 2 ┆ 17.0 │
|
89
|
+
# # │ c ┆ 3 ┆ 1.0 │
|
90
|
+
# # │ b ┆ 5 ┆ 10.0 │
|
91
|
+
# # └─────┴───────┴────────────────┘
|
92
|
+
#
|
93
|
+
# @example Use keyword arguments to easily name your expression inputs.
|
94
|
+
# ldf.group_by("a").agg(
|
95
|
+
# b_sum: Polars.sum("b"),
|
96
|
+
# c_mean_squared: (Polars.col("c") ** 2).mean
|
97
|
+
# ).collect
|
98
|
+
# # =>
|
99
|
+
# # shape: (3, 3)
|
100
|
+
# # ┌─────┬───────┬────────────────┐
|
101
|
+
# # │ a ┆ b_sum ┆ c_mean_squared │
|
102
|
+
# # │ --- ┆ --- ┆ --- │
|
103
|
+
# # │ str ┆ i64 ┆ f64 │
|
104
|
+
# # ╞═════╪═══════╪════════════════╡
|
105
|
+
# # │ a ┆ 2 ┆ 17.0 │
|
106
|
+
# # │ c ┆ 3 ┆ 1.0 │
|
107
|
+
# # │ b ┆ 5 ┆ 10.0 │
|
108
|
+
# # └─────┴───────┴────────────────┘
|
109
|
+
def agg(*aggs, **named_aggs)
|
110
|
+
rbexprs = Utils.parse_as_list_of_expressions(*aggs, **named_aggs)
|
14
111
|
Utils.wrap_ldf(@lgb.agg(rbexprs))
|
15
112
|
end
|
16
113
|
|
@@ -25,12 +25,12 @@ module Polars
|
|
25
25
|
@check_sorted = check_sorted
|
26
26
|
end
|
27
27
|
|
28
|
-
def agg(aggs)
|
28
|
+
def agg(*aggs, **named_aggs)
|
29
29
|
@df.lazy
|
30
30
|
.group_by_rolling(
|
31
31
|
index_column: @time_column, period: @period, offset: @offset, closed: @closed, by: @by, check_sorted: @check_sorted
|
32
32
|
)
|
33
|
-
.agg(aggs)
|
33
|
+
.agg(*aggs, **named_aggs)
|
34
34
|
.collect(no_optimization: true, string_cache: false)
|
35
35
|
end
|
36
36
|
end
|
data/lib/polars/series.rb
CHANGED
@@ -1594,7 +1594,7 @@ module Polars
|
|
1594
1594
|
# # 4
|
1595
1595
|
# # 3
|
1596
1596
|
# # ]
|
1597
|
-
def top_k(k: 5)
|
1597
|
+
def top_k(k: 5, nulls_last: false, multithreaded: true)
|
1598
1598
|
super
|
1599
1599
|
end
|
1600
1600
|
|
@@ -1616,7 +1616,7 @@ module Polars
|
|
1616
1616
|
# # 2
|
1617
1617
|
# # 3
|
1618
1618
|
# # ]
|
1619
|
-
def bottom_k(k: 5)
|
1619
|
+
def bottom_k(k: 5, nulls_last: false, multithreaded: true)
|
1620
1620
|
super
|
1621
1621
|
end
|
1622
1622
|
|
data/lib/polars/string_expr.rb
CHANGED
@@ -840,6 +840,7 @@ module Polars
|
|
840
840
|
# # │ true │
|
841
841
|
# # └──────────┘
|
842
842
|
def json_path_match(json_path)
|
843
|
+
json_path = Utils.parse_as_expression(json_path, str_as_lit: true)
|
843
844
|
Utils.wrap_expr(_rbexpr.str_json_path_match(json_path))
|
844
845
|
end
|
845
846
|
|
@@ -1018,15 +1019,15 @@ module Polars
|
|
1018
1019
|
# )
|
1019
1020
|
# # =>
|
1020
1021
|
# # shape: (3, 3)
|
1021
|
-
# #
|
1022
|
-
# # │ url
|
1023
|
-
# # │ ---
|
1024
|
-
# # │ str
|
1025
|
-
# #
|
1026
|
-
# # │ http://vote.com/ballon_dor?
|
1027
|
-
# # │ http://vote.com/ballon_dor?
|
1028
|
-
# # │ http://vote.com/ballon_dor?
|
1029
|
-
# #
|
1022
|
+
# # ┌─────────────────────────────────┬───────────────────────┬──────────┐
|
1023
|
+
# # │ url ┆ captures ┆ name │
|
1024
|
+
# # │ --- ┆ --- ┆ --- │
|
1025
|
+
# # │ str ┆ struct[2] ┆ str │
|
1026
|
+
# # ╞═════════════════════════════════╪═══════════════════════╪══════════╡
|
1027
|
+
# # │ http://vote.com/ballon_dor?can… ┆ {"messi","python"} ┆ MESSI │
|
1028
|
+
# # │ http://vote.com/ballon_dor?can… ┆ {"weghorst","polars"} ┆ WEGHORST │
|
1029
|
+
# # │ http://vote.com/ballon_dor?err… ┆ {null,null} ┆ null │
|
1030
|
+
# # └─────────────────────────────────┴───────────────────────┴──────────┘
|
1030
1031
|
def extract_groups(pattern)
|
1031
1032
|
Utils.wrap_expr(_rbexpr.str_extract_groups(pattern))
|
1032
1033
|
end
|
@@ -1418,15 +1419,15 @@ module Polars
|
|
1418
1419
|
# )
|
1419
1420
|
# # =>
|
1420
1421
|
# # shape: (3, 2)
|
1421
|
-
# #
|
1422
|
-
# # │ lyrics
|
1423
|
-
# # │ ---
|
1424
|
-
# # │ str
|
1425
|
-
# #
|
1426
|
-
# # │ Everybody wants to rule the
|
1427
|
-
# # │ Tell me what you want, what
|
1428
|
-
# # │ Can you feel the love tonight
|
1429
|
-
# #
|
1422
|
+
# # ┌─────────────────────────────────┬──────────────┐
|
1423
|
+
# # │ lyrics ┆ contains_any │
|
1424
|
+
# # │ --- ┆ --- │
|
1425
|
+
# # │ str ┆ bool │
|
1426
|
+
# # ╞═════════════════════════════════╪══════════════╡
|
1427
|
+
# # │ Everybody wants to rule the wo… ┆ false │
|
1428
|
+
# # │ Tell me what you want, what yo… ┆ true │
|
1429
|
+
# # │ Can you feel the love tonight ┆ true │
|
1430
|
+
# # └─────────────────────────────────┴──────────────┘
|
1430
1431
|
def contains_any(patterns, ascii_case_insensitive: false)
|
1431
1432
|
patterns = Utils.parse_as_expression(patterns, str_as_lit: false, list_as_lit: false)
|
1432
1433
|
Utils.wrap_expr(
|
@@ -1468,15 +1469,15 @@ module Polars
|
|
1468
1469
|
# )
|
1469
1470
|
# # =>
|
1470
1471
|
# # shape: (3, 2)
|
1471
|
-
# #
|
1472
|
-
# # │ lyrics
|
1473
|
-
# # │ ---
|
1474
|
-
# # │ str
|
1475
|
-
# #
|
1476
|
-
# # │ Everybody wants to rule the
|
1477
|
-
# # │ Tell me what you want, what
|
1478
|
-
# # │ Can you feel the love tonight
|
1479
|
-
# #
|
1472
|
+
# # ┌─────────────────────────────────┬─────────────────────────────────┐
|
1473
|
+
# # │ lyrics ┆ removes_pronouns │
|
1474
|
+
# # │ --- ┆ --- │
|
1475
|
+
# # │ str ┆ str │
|
1476
|
+
# # ╞═════════════════════════════════╪═════════════════════════════════╡
|
1477
|
+
# # │ Everybody wants to rule the wo… ┆ Everybody wants to rule the wo… │
|
1478
|
+
# # │ Tell me what you want, what yo… ┆ Tell what want, what really… │
|
1479
|
+
# # │ Can you feel the love tonight ┆ Can feel the love tonight │
|
1480
|
+
# # └─────────────────────────────────┴─────────────────────────────────┘
|
1480
1481
|
#
|
1481
1482
|
# @example
|
1482
1483
|
# df.with_columns(
|
@@ -1489,15 +1490,15 @@ module Polars
|
|
1489
1490
|
# )
|
1490
1491
|
# # =>
|
1491
1492
|
# # shape: (3, 2)
|
1492
|
-
# #
|
1493
|
-
# # │ lyrics
|
1494
|
-
# # │ ---
|
1495
|
-
# # │ str
|
1496
|
-
# #
|
1497
|
-
# # │ Everybody wants to rule the
|
1498
|
-
# # │ Tell me what you want, what
|
1499
|
-
# # │ Can you feel the love tonight
|
1500
|
-
# #
|
1493
|
+
# # ┌─────────────────────────────────┬─────────────────────────────────┐
|
1494
|
+
# # │ lyrics ┆ confusing │
|
1495
|
+
# # │ --- ┆ --- │
|
1496
|
+
# # │ str ┆ str │
|
1497
|
+
# # ╞═════════════════════════════════╪═════════════════════════════════╡
|
1498
|
+
# # │ Everybody wants to rule the wo… ┆ Everybody wants to rule the wo… │
|
1499
|
+
# # │ Tell me what you want, what yo… ┆ Tell you what me want, what me… │
|
1500
|
+
# # │ Can you feel the love tonight ┆ Can me feel the love tonight │
|
1501
|
+
# # └─────────────────────────────────┴─────────────────────────────────┘
|
1501
1502
|
def replace_many(patterns, replace_with, ascii_case_insensitive: false)
|
1502
1503
|
patterns = Utils.parse_as_expression(patterns, str_as_lit: false, list_as_lit: false)
|
1503
1504
|
replace_with = Utils.parse_as_expression(
|
data/lib/polars/utils.rb
CHANGED
@@ -139,7 +139,7 @@ module Polars
|
|
139
139
|
Polars.lit(value)
|
140
140
|
end
|
141
141
|
|
142
|
-
def self.
|
142
|
+
def self.normalize_filepath(path, check_not_directory: true)
|
143
143
|
path = File.expand_path(path)
|
144
144
|
if check_not_directory && File.exist?(path) && Dir.exist?(path)
|
145
145
|
raise ArgumentError, "Expected a file path; #{path} is a directory"
|
@@ -418,5 +418,39 @@ module Polars
|
|
418
418
|
def self.parse_when_inputs(*predicates, **constraints)
|
419
419
|
parse_predicates_constraints_as_expression(*predicates, **constraints)
|
420
420
|
end
|
421
|
+
|
422
|
+
def self.parse_interval_argument(interval)
|
423
|
+
if interval.include?(" ")
|
424
|
+
interval = interval.gsub(" ", "")
|
425
|
+
end
|
426
|
+
interval.downcase
|
427
|
+
end
|
428
|
+
|
429
|
+
def self.validate_rolling_by_aggs_arguments(weights, center:)
|
430
|
+
if !weights.nil?
|
431
|
+
msg = "`weights` is not supported in `rolling_*(..., by=...)` expression"
|
432
|
+
raise InvalidOperationError, msg
|
433
|
+
end
|
434
|
+
if center
|
435
|
+
msg = "`center=True` is not supported in `rolling_*(..., by=...)` expression"
|
436
|
+
raise InvalidOperationError, msg
|
437
|
+
end
|
438
|
+
end
|
439
|
+
|
440
|
+
def self.validate_rolling_aggs_arguments(window_size, closed)
|
441
|
+
if window_size.is_a?(::String)
|
442
|
+
begin
|
443
|
+
window_size = window_size.delete_suffix("i").to_i
|
444
|
+
rescue
|
445
|
+
msg = "Expected a string of the form 'ni', where `n` is a positive integer, got: #{window_size}"
|
446
|
+
raise InvalidOperationError, msg
|
447
|
+
end
|
448
|
+
end
|
449
|
+
if !closed.nil?
|
450
|
+
msg = "`closed` is not supported in `rolling_*(...)` expression"
|
451
|
+
raise InvalidOperationError, msg
|
452
|
+
end
|
453
|
+
window_size
|
454
|
+
end
|
421
455
|
end
|
422
456
|
end
|
data/lib/polars/version.rb
CHANGED
data/lib/polars.rb
CHANGED
@@ -42,9 +42,17 @@ require_relative "polars/functions/whenthen"
|
|
42
42
|
require_relative "polars/functions/aggregation/horizontal"
|
43
43
|
require_relative "polars/functions/aggregation/vertical"
|
44
44
|
require_relative "polars/functions/range/date_range"
|
45
|
+
require_relative "polars/functions/range/datetime_range"
|
45
46
|
require_relative "polars/functions/range/int_range"
|
47
|
+
require_relative "polars/functions/range/time_range"
|
46
48
|
require_relative "polars/group_by"
|
47
|
-
require_relative "polars/io"
|
49
|
+
require_relative "polars/io/avro"
|
50
|
+
require_relative "polars/io/csv"
|
51
|
+
require_relative "polars/io/database"
|
52
|
+
require_relative "polars/io/ipc"
|
53
|
+
require_relative "polars/io/json"
|
54
|
+
require_relative "polars/io/ndjson"
|
55
|
+
require_relative "polars/io/parquet"
|
48
56
|
require_relative "polars/lazy_frame"
|
49
57
|
require_relative "polars/lazy_group_by"
|
50
58
|
require_relative "polars/list_expr"
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: polars-df
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.11.0
|
5
5
|
platform: x86_64-linux-musl
|
6
6
|
authors:
|
7
7
|
- Andrew Kane
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2024-
|
11
|
+
date: 2024-06-03 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: bigdecimal
|
@@ -71,11 +71,19 @@ files:
|
|
71
71
|
- lib/polars/functions/lit.rb
|
72
72
|
- lib/polars/functions/random.rb
|
73
73
|
- lib/polars/functions/range/date_range.rb
|
74
|
+
- lib/polars/functions/range/datetime_range.rb
|
74
75
|
- lib/polars/functions/range/int_range.rb
|
76
|
+
- lib/polars/functions/range/time_range.rb
|
75
77
|
- lib/polars/functions/repeat.rb
|
76
78
|
- lib/polars/functions/whenthen.rb
|
77
79
|
- lib/polars/group_by.rb
|
78
|
-
- lib/polars/io.rb
|
80
|
+
- lib/polars/io/avro.rb
|
81
|
+
- lib/polars/io/csv.rb
|
82
|
+
- lib/polars/io/database.rb
|
83
|
+
- lib/polars/io/ipc.rb
|
84
|
+
- lib/polars/io/json.rb
|
85
|
+
- lib/polars/io/ndjson.rb
|
86
|
+
- lib/polars/io/parquet.rb
|
79
87
|
- lib/polars/lazy_frame.rb
|
80
88
|
- lib/polars/lazy_group_by.rb
|
81
89
|
- lib/polars/list_expr.rb
|
@@ -96,7 +104,7 @@ files:
|
|
96
104
|
- lib/polars/utils.rb
|
97
105
|
- lib/polars/version.rb
|
98
106
|
- lib/polars/whenthen.rb
|
99
|
-
homepage: https://github.com/ankane/polars
|
107
|
+
homepage: https://github.com/ankane/ruby-polars
|
100
108
|
licenses:
|
101
109
|
- MIT
|
102
110
|
metadata: {}
|