polars-df 0.1.2 → 0.1.3
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.yardopts +3 -0
- data/CHANGELOG.md +4 -0
- data/Cargo.lock +2 -1
- data/README.md +1 -1
- data/ext/polars/Cargo.toml +7 -1
- data/ext/polars/src/conversion.rs +35 -2
- data/ext/polars/src/dataframe.rs +228 -11
- data/ext/polars/src/lazy/dataframe.rs +3 -3
- data/ext/polars/src/lazy/dsl.rs +59 -2
- data/ext/polars/src/lib.rs +151 -10
- data/ext/polars/src/series.rs +182 -29
- data/ext/polars/src/set.rs +91 -0
- data/ext/polars/src/utils.rs +19 -0
- data/lib/polars/batched_csv_reader.rb +1 -0
- data/lib/polars/cat_expr.rb +39 -0
- data/lib/polars/data_frame.rb +2284 -137
- data/lib/polars/date_time_expr.rb +1282 -7
- data/lib/polars/exceptions.rb +20 -0
- data/lib/polars/expr.rb +612 -7
- data/lib/polars/expr_dispatch.rb +14 -0
- data/lib/polars/functions.rb +219 -0
- data/lib/polars/group_by.rb +517 -0
- data/lib/polars/io.rb +421 -2
- data/lib/polars/lazy_frame.rb +1261 -67
- data/lib/polars/lazy_functions.rb +288 -10
- data/lib/polars/lazy_group_by.rb +79 -0
- data/lib/polars/list_expr.rb +5 -0
- data/lib/polars/meta_expr.rb +21 -0
- data/lib/polars/series.rb +1476 -212
- data/lib/polars/slice.rb +104 -0
- data/lib/polars/string_expr.rb +663 -2
- data/lib/polars/struct_expr.rb +73 -0
- data/lib/polars/utils.rb +43 -3
- data/lib/polars/version.rb +2 -1
- data/lib/polars/when.rb +1 -0
- data/lib/polars/when_then.rb +1 -0
- data/lib/polars.rb +7 -10
- metadata +9 -2
@@ -0,0 +1,14 @@
|
|
1
|
+
module Polars
|
2
|
+
# @private
|
3
|
+
module ExprDispatch
|
4
|
+
private
|
5
|
+
|
6
|
+
def method_missing(method, ...)
|
7
|
+
return super unless self.class.method_defined?(method)
|
8
|
+
|
9
|
+
s = Utils.wrap_s(_s)
|
10
|
+
expr = Utils.col(s.name)
|
11
|
+
s.to_frame.select(expr.send(method, ...)).to_series
|
12
|
+
end
|
13
|
+
end
|
14
|
+
end
|
data/lib/polars/functions.rb
CHANGED
@@ -1,5 +1,51 @@
|
|
1
1
|
module Polars
|
2
2
|
module Functions
|
3
|
+
# Convert categorical variables into dummy/indicator variables.
|
4
|
+
#
|
5
|
+
# @param df [DataFrame]
|
6
|
+
# DataFrame to convert.
|
7
|
+
# @param columns [Array, nil]
|
8
|
+
# A subset of columns to convert to dummy variables. `nil` means
|
9
|
+
# "all columns".
|
10
|
+
#
|
11
|
+
# @return [DataFrame]
|
12
|
+
def get_dummies(df, columns: nil)
|
13
|
+
df.to_dummies(columns: columns)
|
14
|
+
end
|
15
|
+
|
16
|
+
# Aggregate multiple Dataframes/Series to a single DataFrame/Series.
|
17
|
+
#
|
18
|
+
# @param items [Object]
|
19
|
+
# DataFrames/Series/LazyFrames to concatenate.
|
20
|
+
# @param rechunk [Boolean]
|
21
|
+
# Make sure that all data is in contiguous memory.
|
22
|
+
# @param how ["vertical", "diagonal", "horizontal"]
|
23
|
+
# Lazy only supports the 'vertical' strategy.
|
24
|
+
#
|
25
|
+
# - Vertical: applies multiple `vstack` operations.
|
26
|
+
# - Diagonal: finds a union between the column schemas and fills missing column values with null.
|
27
|
+
# - Horizontal: stacks Series horizontally and fills with nulls if the lengths don't match.
|
28
|
+
# @param parallel [Boolean]
|
29
|
+
# Only relevant for LazyFrames. This determines if the concatenated
|
30
|
+
# lazy computations may be executed in parallel.
|
31
|
+
#
|
32
|
+
# @return [Object]
|
33
|
+
#
|
34
|
+
# @example
|
35
|
+
# df1 = Polars::DataFrame.new({"a" => [1], "b" => [3]})
|
36
|
+
# df2 = Polars::DataFrame.new({"a" => [2], "b" => [4]})
|
37
|
+
# Polars.concat([df1, df2])
|
38
|
+
# # =>
|
39
|
+
# # shape: (2, 2)
|
40
|
+
# # ┌─────┬─────┐
|
41
|
+
# # │ a ┆ b │
|
42
|
+
# # │ --- ┆ --- │
|
43
|
+
# # │ i64 ┆ i64 │
|
44
|
+
# # ╞═════╪═════╡
|
45
|
+
# # │ 1 ┆ 3 │
|
46
|
+
# # ├╌╌╌╌╌┼╌╌╌╌╌┤
|
47
|
+
# # │ 2 ┆ 4 │
|
48
|
+
# # └─────┴─────┘
|
3
49
|
def concat(items, rechunk: true, how: "vertical", parallel: true)
|
4
50
|
if items.empty?
|
5
51
|
raise ArgumentError, "cannot concat empty list"
|
@@ -41,5 +87,178 @@ module Polars
|
|
41
87
|
out
|
42
88
|
end
|
43
89
|
end
|
90
|
+
|
91
|
+
# Create a range of type `Datetime` (or `Date`).
|
92
|
+
#
|
93
|
+
# @param low [Object]
|
94
|
+
# Lower bound of the date range.
|
95
|
+
# @param high [Object]
|
96
|
+
# Upper bound of the date range.
|
97
|
+
# @param interval [Object]
|
98
|
+
# Interval periods. It can be a polars duration string, such as `3d12h4m25s`
|
99
|
+
# representing 3 days, 12 hours, 4 minutes, and 25 seconds.
|
100
|
+
# @param lazy [Boolean]
|
101
|
+
# Return an expression.
|
102
|
+
# @param closed ["both", "left", "right", "none"]
|
103
|
+
# Define whether the temporal window interval is closed or not.
|
104
|
+
# @param name [String]
|
105
|
+
# Name of the output Series.
|
106
|
+
# @param time_unit [nil, "ns", "us", "ms"]
|
107
|
+
# Set the time unit.
|
108
|
+
# @param time_zone [String]
|
109
|
+
# Optional timezone
|
110
|
+
#
|
111
|
+
# @return [Object]
|
112
|
+
#
|
113
|
+
# @note
|
114
|
+
# If both `low` and `high` are passed as date types (not datetime), and the
|
115
|
+
# interval granularity is no finer than 1d, the returned range is also of
|
116
|
+
# type date. All other permutations return a datetime Series.
|
117
|
+
#
|
118
|
+
# @example Using polars duration string to specify the interval
|
119
|
+
# Polars.date_range(Date.new(2022, 1, 1), Date.new(2022, 3, 1), "1mo", name: "drange")
|
120
|
+
# # =>
|
121
|
+
# # shape: (3,)
|
122
|
+
# # Series: 'drange' [date]
|
123
|
+
# # [
|
124
|
+
# # 2022-01-01
|
125
|
+
# # 2022-02-01
|
126
|
+
# # 2022-03-01
|
127
|
+
# # ]
|
128
|
+
#
|
129
|
+
# @example Using `timedelta` object to specify the interval:
|
130
|
+
# Polars.date_range(
|
131
|
+
# DateTime.new(1985, 1, 1),
|
132
|
+
# DateTime.new(1985, 1, 10),
|
133
|
+
# "1d12h",
|
134
|
+
# time_unit: "ms"
|
135
|
+
# )
|
136
|
+
# # =>
|
137
|
+
# # shape: (7,)
|
138
|
+
# # Series: '' [datetime[ms]]
|
139
|
+
# # [
|
140
|
+
# # 1985-01-01 00:00:00
|
141
|
+
# # 1985-01-02 12:00:00
|
142
|
+
# # 1985-01-04 00:00:00
|
143
|
+
# # 1985-01-05 12:00:00
|
144
|
+
# # 1985-01-07 00:00:00
|
145
|
+
# # 1985-01-08 12:00:00
|
146
|
+
# # 1985-01-10 00:00:00
|
147
|
+
# # ]
|
148
|
+
def date_range(
|
149
|
+
low,
|
150
|
+
high,
|
151
|
+
interval,
|
152
|
+
lazy: false,
|
153
|
+
closed: "both",
|
154
|
+
name: nil,
|
155
|
+
time_unit: nil,
|
156
|
+
time_zone: nil
|
157
|
+
)
|
158
|
+
if defined?(ActiveSupport::Duration) && interval.is_a?(ActiveSupport::Duration)
|
159
|
+
raise Todo
|
160
|
+
else
|
161
|
+
interval = interval.to_s
|
162
|
+
if interval.include?(" ")
|
163
|
+
interval = interval.gsub(" ", "")
|
164
|
+
end
|
165
|
+
end
|
166
|
+
|
167
|
+
if low.is_a?(Expr) || high.is_a?(Expr) || lazy
|
168
|
+
low = Utils.expr_to_lit_or_expr(low, str_to_lit: true)
|
169
|
+
high = Utils.expr_to_lit_or_expr(high, str_to_lit: true)
|
170
|
+
return Utils.wrap_expr(
|
171
|
+
_rb_date_range_lazy(low, high, interval, closed, name, time_zone)
|
172
|
+
)
|
173
|
+
end
|
174
|
+
|
175
|
+
low, low_is_date = _ensure_datetime(low)
|
176
|
+
high, high_is_date = _ensure_datetime(high)
|
177
|
+
|
178
|
+
if !time_unit.nil?
|
179
|
+
tu = time_unit
|
180
|
+
elsif interval.include?("ns")
|
181
|
+
tu = "ns"
|
182
|
+
else
|
183
|
+
tu = "us"
|
184
|
+
end
|
185
|
+
|
186
|
+
start = Utils._datetime_to_pl_timestamp(low, tu)
|
187
|
+
stop = Utils._datetime_to_pl_timestamp(high, tu)
|
188
|
+
if name.nil?
|
189
|
+
name = ""
|
190
|
+
end
|
191
|
+
|
192
|
+
dt_range = Utils.wrap_s(
|
193
|
+
_rb_date_range(start, stop, interval, closed, name, tu, time_zone)
|
194
|
+
)
|
195
|
+
if low_is_date && high_is_date && !["h", "m", "s"].any? { |v| _interval_granularity(interval).end_with?(v) }
|
196
|
+
dt_range = dt_range.cast(Date)
|
197
|
+
end
|
198
|
+
|
199
|
+
dt_range
|
200
|
+
end
|
201
|
+
|
202
|
+
# def cut
|
203
|
+
# end
|
204
|
+
|
205
|
+
# def align_frames
|
206
|
+
# end
|
207
|
+
|
208
|
+
# Return a new Series of given length and type, filled with ones.
|
209
|
+
#
|
210
|
+
# @param n [Integer]
|
211
|
+
# Number of elements in the `Series`
|
212
|
+
# @param dtype [Symbol]
|
213
|
+
# DataType of the elements, defaults to `:f64`
|
214
|
+
#
|
215
|
+
# @return [Series]
|
216
|
+
#
|
217
|
+
# @note
|
218
|
+
# In the lazy API you should probably not use this, but use `lit(1)`
|
219
|
+
# instead.
|
220
|
+
def ones(n, dtype: nil)
|
221
|
+
s = Series.new([1.0])
|
222
|
+
if dtype
|
223
|
+
s = s.cast(dtype)
|
224
|
+
end
|
225
|
+
s.new_from_index(0, n)
|
226
|
+
end
|
227
|
+
|
228
|
+
# Return a new Series of given length and type, filled with zeros.
|
229
|
+
#
|
230
|
+
# @param n [Integer]
|
231
|
+
# Number of elements in the `Series`
|
232
|
+
# @param dtype [Symbol]
|
233
|
+
# DataType of the elements, defaults to `:f64`
|
234
|
+
#
|
235
|
+
# @return [Series]
|
236
|
+
#
|
237
|
+
# @note
|
238
|
+
# In the lazy API you should probably not use this, but use `lit(0)`
|
239
|
+
# instead.
|
240
|
+
def zeros(n, dtype: nil)
|
241
|
+
s = Series.new([0.0])
|
242
|
+
if dtype
|
243
|
+
s = s.cast(dtype)
|
244
|
+
end
|
245
|
+
s.new_from_index(0, n)
|
246
|
+
end
|
247
|
+
|
248
|
+
private
|
249
|
+
|
250
|
+
def _ensure_datetime(value)
|
251
|
+
is_date_type = false
|
252
|
+
if !value.is_a?(DateTime)
|
253
|
+
value = DateTime.new(value.year, value.month, value.day)
|
254
|
+
is_date_type = true
|
255
|
+
end
|
256
|
+
[value, is_date_type]
|
257
|
+
end
|
258
|
+
|
259
|
+
# TODO
|
260
|
+
def _interval_granularity(interval)
|
261
|
+
interval
|
262
|
+
end
|
44
263
|
end
|
45
264
|
end
|