polars-df 0.1.1 → 0.1.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +4 -0
- data/Cargo.lock +1 -1
- data/ext/polars/Cargo.toml +1 -1
- data/ext/polars/src/batched_csv.rs +120 -0
- data/ext/polars/src/conversion.rs +105 -5
- data/ext/polars/src/dataframe.rs +132 -4
- data/ext/polars/src/error.rs +9 -0
- data/ext/polars/src/file.rs +8 -7
- data/ext/polars/src/lazy/apply.rs +7 -0
- data/ext/polars/src/lazy/dataframe.rs +132 -0
- data/ext/polars/src/lazy/dsl.rs +38 -0
- data/ext/polars/src/lazy/meta.rs +1 -1
- data/ext/polars/src/lazy/mod.rs +1 -0
- data/ext/polars/src/lib.rs +77 -3
- data/ext/polars/src/series.rs +8 -9
- data/lib/polars/batched_csv_reader.rb +95 -0
- data/lib/polars/data_frame.rb +585 -19
- data/lib/polars/expr.rb +17 -2
- data/lib/polars/io.rb +342 -2
- data/lib/polars/lazy_frame.rb +156 -2
- data/lib/polars/lazy_functions.rb +154 -11
- data/lib/polars/series.rb +806 -18
- data/lib/polars/utils.rb +33 -0
- data/lib/polars/version.rb +1 -1
- data/lib/polars.rb +9 -0
- metadata +5 -2
@@ -1,14 +1,45 @@
|
|
1
1
|
module Polars
|
2
2
|
module LazyFunctions
|
3
3
|
def col(name)
|
4
|
-
|
5
|
-
|
4
|
+
if name.is_a?(Series)
|
5
|
+
name = name.to_a
|
6
|
+
end
|
7
|
+
|
8
|
+
if name.is_a?(Array)
|
9
|
+
if name.length == 0 || name[0].is_a?(String) || name[0].is_a?(Symbol)
|
10
|
+
name = name.map { |v| v.is_a?(Symbol) ? v.to_s : v }
|
11
|
+
Utils.wrap_expr(RbExpr.cols(name))
|
12
|
+
elsif Utils.is_polars_dtype(name[0])
|
13
|
+
raise Todo
|
14
|
+
# Utils.wrap_expr(_dtype_cols(name))
|
15
|
+
else
|
16
|
+
raise ArgumentError, "Expected list values to be all `str` or all `DataType`"
|
17
|
+
end
|
18
|
+
else
|
19
|
+
name = name.to_s if name.is_a?(Symbol)
|
20
|
+
Utils.wrap_expr(RbExpr.col(name))
|
21
|
+
end
|
6
22
|
end
|
7
23
|
|
8
24
|
def element
|
9
25
|
col("")
|
10
26
|
end
|
11
27
|
|
28
|
+
def count(column = nil)
|
29
|
+
if column.nil?
|
30
|
+
return Utils.wrap_expr(RbExpr.count)
|
31
|
+
end
|
32
|
+
|
33
|
+
if column.is_a?(Series)
|
34
|
+
column.len
|
35
|
+
else
|
36
|
+
col(column).count
|
37
|
+
end
|
38
|
+
end
|
39
|
+
|
40
|
+
# def to_list
|
41
|
+
# end
|
42
|
+
|
12
43
|
def std(column, ddof: 1)
|
13
44
|
if column.is_a?(Series)
|
14
45
|
column.std(ddof: ddof)
|
@@ -59,7 +90,7 @@ module Polars
|
|
59
90
|
# TODO
|
60
91
|
Utils.wrap_expr(_sum_exprs(exprs))
|
61
92
|
else
|
62
|
-
raise
|
93
|
+
raise Todo
|
63
94
|
end
|
64
95
|
end
|
65
96
|
|
@@ -83,10 +114,97 @@ module Polars
|
|
83
114
|
end
|
84
115
|
end
|
85
116
|
|
117
|
+
# def n_unique
|
118
|
+
# end
|
119
|
+
|
120
|
+
def first(column = nil)
|
121
|
+
if column.nil?
|
122
|
+
return Utils.wrap_expr(RbExpr.first)
|
123
|
+
end
|
124
|
+
|
125
|
+
if column.is_a?(Series)
|
126
|
+
if column.len > 0
|
127
|
+
column[0]
|
128
|
+
else
|
129
|
+
raise IndexError, "The series is empty, so no first value can be returned."
|
130
|
+
end
|
131
|
+
else
|
132
|
+
col(column).first
|
133
|
+
end
|
134
|
+
end
|
135
|
+
|
136
|
+
# def last
|
137
|
+
# end
|
138
|
+
|
139
|
+
# def head
|
140
|
+
# end
|
141
|
+
|
142
|
+
# def tail
|
143
|
+
# end
|
144
|
+
|
86
145
|
def lit(value)
|
87
146
|
Utils.wrap_expr(RbExpr.lit(value))
|
88
147
|
end
|
89
148
|
|
149
|
+
# def cumsum
|
150
|
+
# end
|
151
|
+
|
152
|
+
# def spearman_rank_corr
|
153
|
+
# end
|
154
|
+
|
155
|
+
# def pearson_corr
|
156
|
+
# end
|
157
|
+
|
158
|
+
# def cov
|
159
|
+
# end
|
160
|
+
|
161
|
+
# def map
|
162
|
+
# end
|
163
|
+
|
164
|
+
# def apply
|
165
|
+
# end
|
166
|
+
|
167
|
+
def fold(acc, f, exprs)
|
168
|
+
acc = Utils.expr_to_lit_or_expr(acc, str_to_lit: true)
|
169
|
+
if exprs.is_a?(Expr)
|
170
|
+
exprs = [exprs]
|
171
|
+
end
|
172
|
+
|
173
|
+
exprs = Utils.selection_to_rbexpr_list(exprs)
|
174
|
+
Utils.wrap_expr(RbExpr.fold(acc._rbexpr, f, exprs))
|
175
|
+
end
|
176
|
+
|
177
|
+
# def reduce
|
178
|
+
# end
|
179
|
+
|
180
|
+
# def cumfold
|
181
|
+
# end
|
182
|
+
|
183
|
+
# def cumreduce
|
184
|
+
# end
|
185
|
+
|
186
|
+
# def any
|
187
|
+
# end
|
188
|
+
|
189
|
+
# def exclude
|
190
|
+
# end
|
191
|
+
|
192
|
+
def all(name = nil)
|
193
|
+
if name.nil?
|
194
|
+
col("*")
|
195
|
+
elsif name.is_a?(String) || name.is_a?(Symbol)
|
196
|
+
col(name).all
|
197
|
+
else
|
198
|
+
raise Todo
|
199
|
+
end
|
200
|
+
end
|
201
|
+
|
202
|
+
# def groups
|
203
|
+
# end
|
204
|
+
|
205
|
+
# def quantile
|
206
|
+
# end
|
207
|
+
|
90
208
|
def arange(low, high, step: 1, eager: false, dtype: nil)
|
91
209
|
low = Utils.expr_to_lit_or_expr(low, str_to_lit: false)
|
92
210
|
high = Utils.expr_to_lit_or_expr(high, str_to_lit: false)
|
@@ -106,16 +224,41 @@ module Polars
|
|
106
224
|
end
|
107
225
|
end
|
108
226
|
|
109
|
-
def
|
110
|
-
|
111
|
-
|
112
|
-
|
113
|
-
|
114
|
-
|
115
|
-
|
116
|
-
|
227
|
+
# def argsort_by
|
228
|
+
# end
|
229
|
+
|
230
|
+
# def duration
|
231
|
+
# end
|
232
|
+
|
233
|
+
# def format
|
234
|
+
# end
|
235
|
+
|
236
|
+
def concat_list(exprs)
|
237
|
+
exprs = Utils.selection_to_rbexpr_list(exprs)
|
238
|
+
Utils.wrap_expr(RbExpr.concat_lst(exprs))
|
117
239
|
end
|
118
240
|
|
241
|
+
# def collect_all
|
242
|
+
# end
|
243
|
+
|
244
|
+
# def select
|
245
|
+
# end
|
246
|
+
|
247
|
+
# def struct
|
248
|
+
# end
|
249
|
+
|
250
|
+
# def repeat
|
251
|
+
# end
|
252
|
+
|
253
|
+
# def arg_where
|
254
|
+
# end
|
255
|
+
|
256
|
+
# def coalesce
|
257
|
+
# end
|
258
|
+
|
259
|
+
# def from_epoch
|
260
|
+
# end
|
261
|
+
|
119
262
|
def when(expr)
|
120
263
|
expr = Utils.expr_to_lit_or_expr(expr)
|
121
264
|
pw = RbExpr.when(expr._rbexpr)
|