Book a Demo!
CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutPoliciesSign UpSign In
polakowo
GitHub Repository: polakowo/vectorbt
 Path: blob/master/tests/notebooks/records.ipynb
1073 views
Kernel: Python 3 (ipykernel)

records

import vectorbt as vbt

import numpy as np
import pandas as pd
from numba import njit
from collections import namedtuple
from datetime import datetime

# Disable caching for performance testing
# NOTE: Expect waterfall of executions, since some attributes depend on other attributes 
# that have to be calculated again and again
vbt.settings.caching['enabled'] = False

example_dt = np.dtype([
    ('id', np.int64),
    ('idx', np.int64),
    ('col', np.int64),
    ('some_field1', np.float64),
    ('some_field2', np.float64)
], align=True)

records_arr = np.asarray([
    (0, 0, 0, 10, 21),
    (1, 1, 0, 11, 20),
    (2, 2, 0, 12, 19),
    (3, 0, 1, 13, 18),
    (4, 1, 1, 14, 17),
    (5, 2, 1, 13, 18),
    (6, 0, 2, 12, 19),
    (7, 1, 2, 11, 20),
    (8, 2, 2, 10, 21)
], dtype=example_dt)
print(records_arr)
print(records_arr.shape)

columns = pd.MultiIndex.from_arrays([[0, 1, 1, 1], ['a', 'b', 'c', 'd']], names=['lvl1', 'lvl2'])
wrapper = vbt.ArrayWrapper(index=['x', 'y', 'z'], columns=columns, ndim=2, freq='1 days')
records = vbt.Records(wrapper, records_arr)

records_grouped = vbt.Records(wrapper.copy(group_by=0), records_arr)
[(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.) (7, 1, 2, 11., 20.) (8, 2, 2, 10., 21.)] (9,) 
big_records_arr = np.asarray(list(zip(*(
    np.arange(1000000),
    np.tile(np.arange(1000), 1000),
    np.repeat(np.arange(1000), 1000),
    np.random.randint(0, 100, size=1000000),
    np.random.randint(0, 100, size=1000000)))), dtype=example_dt)
print(big_records_arr.shape)

big_columns = pd.MultiIndex.from_arrays([np.repeat(np.array([0, 1]), 500), np.arange(1000)], names=['lvl1', 'lvl2'])
big_wrapper = vbt.ArrayWrapper(index=np.arange(1000), columns=big_columns, ndim=2, freq='1 days')
big_records = vbt.Records(big_wrapper, big_records_arr)

big_records_grouped = vbt.Records(big_wrapper.copy(group_by=0), big_records_arr)
(1000000,) 
records_nosort = records.copy(records_arr=records.records_arr[::-1])
print(records_nosort.records_arr)

big_records_nosort = big_records.copy(records_arr=big_records.records_arr[::-1])
[(8, 2, 2, 10., 21.) (7, 1, 2, 11., 20.) (6, 0, 2, 12., 19.) (5, 2, 1, 13., 18.) (4, 1, 1, 14., 17.) (3, 0, 1, 13., 18.) (2, 2, 0, 12., 19.) (1, 1, 0, 11., 20.) (0, 0, 0, 10., 21.)] 
group_by = pd.Series(['first', 'first', 'second', 'second'], name='group')
big_group_by = pd.Series(np.repeat(np.array([0, 1]), 500))

ColumnMapper

print(records.col_mapper.col_arr)
[0 0 0 1 1 1 2 2 2] 
print(records.col_mapper.get_col_arr())
print(records_grouped.col_mapper.get_col_arr())

%timeit big_records_grouped.col_mapper.get_col_arr()
[0 0 0 1 1 1 2 2 2] [0 0 0 1 1 1 1 1 1] 7.57 µs ± 81.5 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.col_mapper.col_range)

%timeit big_records.col_mapper.col_range
[[ 0 3] [ 3 6] [ 6 9] [-1 -1]] 6.99 µs ± 73.1 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.col_mapper.get_col_range())
print(records_grouped.col_mapper.get_col_range())

%timeit big_records_grouped.col_mapper.get_col_range()
[[ 0 3] [ 3 6] [ 6 9] [-1 -1]] [[0 3] [3 9]] 7.38 µs ± 56.5 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.col_mapper.col_map)

%timeit big_records.col_mapper.col_map
(array([0, 1, 2, 3, 4, 5, 6, 7, 8]), array([3, 3, 3, 0])) 7.03 µs ± 22.4 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.col_mapper.get_col_map())
print(records_grouped.col_mapper.get_col_map())

%timeit big_records_grouped.col_mapper.get_col_map()
(array([0, 1, 2, 3, 4, 5, 6, 7, 8]), array([3, 3, 3, 0])) (array([0, 1, 2, 3, 4, 5, 6, 7, 8]), array([3, 6])) 7.43 µs ± 46.4 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.col_mapper.is_sorted())
%timeit big_records.col_mapper.is_sorted()

print(records_nosort.col_mapper.is_sorted())
%timeit big_records_nosort.col_mapper.is_sorted()
True 7.48 µs ± 48.2 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) False 7.49 µs ± 8.8 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

MappedArray

mapped_array = records.map_field('some_field1')
big_mapped_array = big_records.map_field('some_field1')

mapped_array_nosort = records_nosort.map_field('some_field1')
big_mapped_array_nosort = big_records_nosort.map_field('some_field1')

mapped_array_grouped = records_grouped.map_field('some_field1')
big_mapped_array_grouped = big_records_grouped.map_field('some_field1')

print(mapped_array[(0, 'a')].values)
print(mapped_array[(0, 'a')].col_arr)
print(mapped_array[(0, 'a')].wrapper.columns)

print(mapped_array[(1, 'b')].values)
print(mapped_array[(1, 'b')].col_arr)
print(mapped_array[(1, 'b')].wrapper.columns)

print(mapped_array[[(0, 'a'), (0, 'a')]].values)
print(mapped_array[[(0, 'a'), (0, 'a')]].col_arr)
print(mapped_array[[(0, 'a'), (0, 'a')]].wrapper.columns)

print(mapped_array[[(0, 'a'), (1, 'b')]].values)
print(mapped_array[[(0, 'a'), (1, 'b')]].col_arr)
print(mapped_array[[(0, 'a'), (1, 'b')]].wrapper.columns)

%timeit big_mapped_array.iloc[0]
%timeit big_mapped_array.iloc[:]
[10. 11. 12.] [0 0 0] MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) [13. 14. 13.] [0 0 0] MultiIndex([(1, 'b')], names=['lvl1', 'lvl2']) [10. 11. 12. 10. 11. 12.] [0 0 0 1 1 1] MultiIndex([(0, 'a'), (0, 'a')], names=['lvl1', 'lvl2']) [10. 11. 12. 13. 14. 13.] [0 0 0 1 1 1] MultiIndex([(0, 'a'), (1, 'b')], names=['lvl1', 'lvl2']) 954 µs ± 11.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 13.7 ms ± 1.09 ms per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array_nosort[(0, 'a')].values)
print(mapped_array_nosort[(0, 'a')].col_arr)
print(mapped_array_nosort[(0, 'a')].wrapper.columns)

print(mapped_array_nosort[(1, 'b')].values)
print(mapped_array_nosort[(1, 'b')].col_arr)
print(mapped_array_nosort[(1, 'b')].wrapper.columns)

print(mapped_array_nosort[[(0, 'a'), (0, 'a')]].values)
print(mapped_array_nosort[[(0, 'a'), (0, 'a')]].col_arr)
print(mapped_array_nosort[[(0, 'a'), (0, 'a')]].wrapper.columns)

print(mapped_array_nosort[[(0, 'a'), (1, 'b')]].values)
print(mapped_array_nosort[[(0, 'a'), (1, 'b')]].col_arr)
print(mapped_array_nosort[[(0, 'a'), (1, 'b')]].wrapper.columns)

%timeit big_mapped_array_nosort.iloc[0]
%timeit big_mapped_array_nosort.iloc[:]
[12. 11. 10.] [0 0 0] MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) [13. 14. 13.] [0 0 0] MultiIndex([(1, 'b')], names=['lvl1', 'lvl2']) [12. 11. 10. 12. 11. 10.] [0 0 0 1 1 1] MultiIndex([(0, 'a'), (0, 'a')], names=['lvl1', 'lvl2']) [12. 11. 10. 13. 14. 13.] [0 0 0 1 1 1] MultiIndex([(0, 'a'), (1, 'b')], names=['lvl1', 'lvl2']) 954 µs ± 7.38 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 14.2 ms ± 448 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array_grouped[0].wrapper.columns)  # indexing on groups, not columns!
print(mapped_array_grouped[0].wrapper.ndim)
print(mapped_array_grouped[0].wrapper.grouped_ndim)
print(mapped_array_grouped[0].wrapper.grouper.group_by)

print(mapped_array_grouped[1].wrapper.columns)
print(mapped_array_grouped[1].wrapper.ndim)
print(mapped_array_grouped[1].wrapper.grouped_ndim)
print(mapped_array_grouped[1].wrapper.grouper.group_by)

print(mapped_array_grouped[[0]].wrapper.columns)
print(mapped_array_grouped[[0]].wrapper.ndim)
print(mapped_array_grouped[[0]].wrapper.grouped_ndim)
print(mapped_array_grouped[[0]].wrapper.grouper.group_by)

print(mapped_array_grouped[[0, 1]].wrapper.columns)
print(mapped_array_grouped[[0, 1]].wrapper.ndim)
print(mapped_array_grouped[[0, 1]].wrapper.grouped_ndim)
print(mapped_array_grouped[[0, 1]].wrapper.grouper.group_by)

%timeit big_mapped_array_grouped.iloc[0]
%timeit big_mapped_array_grouped.iloc[:]
MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) 1 1 Index([0], dtype='int64', name='lvl1') MultiIndex([(1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 1 Index([1, 1, 1], dtype='int64', name='lvl1') MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) 2 2 Index([0], dtype='int64', name='lvl1') MultiIndex([(0, 'a'), (1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 2 Index([0, 1, 1, 1], dtype='int64', name='lvl1') 5.57 ms ± 375 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 12.1 ms ± 451 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.wrapper.index)
print(mapped_array.wrapper.columns)
print(mapped_array.wrapper.ndim)
print(mapped_array.wrapper.grouper.group_by)

print(mapped_array_grouped.wrapper.index)
print(mapped_array_grouped.wrapper.columns)
print(mapped_array_grouped.wrapper.ndim)
print(mapped_array_grouped.wrapper.grouper.group_by)
Index(['x', 'y', 'z'], dtype='object') MultiIndex([(0, 'a'), (1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 None Index(['x', 'y', 'z'], dtype='object') MultiIndex([(0, 'a'), (1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 Index([0, 1, 1, 1], dtype='int64', name='lvl1') 
print(mapped_array.values)
[10. 11. 12. 13. 14. 13. 12. 11. 10.] 
print(mapped_array.col_arr)
[0 0 0 1 1 1 2 2 2] 
print(mapped_array.id_arr)
[0 1 2 3 4 5 6 7 8] 
print(mapped_array.idx_arr)
[0 1 2 0 1 2 0 1 2] 
print(mapped_array.is_sorted())
%timeit big_mapped_array.is_sorted()

print(mapped_array_nosort.is_sorted())
%timeit big_mapped_array_nosort.is_sorted()

print(mapped_array.is_sorted(incl_id=True))
%timeit big_mapped_array.is_sorted(incl_id=True)

print(mapped_array_nosort.is_sorted(incl_id=True))
%timeit big_mapped_array_nosort.is_sorted(incl_id=True)
True 1.04 ms ± 7.48 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) False 7.84 µs ± 31 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) True 1.58 ms ± 65.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) False 6.99 µs ± 20.4 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(mapped_array.sort().col_arr)
print(mapped_array.sort().id_arr)
%timeit big_mapped_array.sort()

print(mapped_array_nosort.sort().col_arr)
print(mapped_array_nosort.sort().id_arr)
%timeit big_mapped_array_nosort.sort()

print(mapped_array.sort(incl_id=True).col_arr)
print(mapped_array.sort(incl_id=True).id_arr)
%timeit big_mapped_array.sort(incl_id=True)

print(mapped_array_nosort.sort(incl_id=True).col_arr)
print(mapped_array_nosort.sort(incl_id=True).id_arr)
%timeit big_mapped_array_nosort.sort(incl_id=True)
[0 0 0 1 1 1 2 2 2] [0 1 2 3 4 5 6 7 8] 1.41 ms ± 48.7 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) [0 0 0 1 1 1 2 2 2] [2 1 0 5 4 3 8 7 6] 20.1 ms ± 544 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) [0 0 0 1 1 1 2 2 2] [0 1 2 3 4 5 6 7 8] 1.91 ms ± 3.77 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) [0 0 0 1 1 1 2 2 2] [0 1 2 3 4 5 6 7 8] 16.1 ms ± 1.42 ms per loop (mean ± std. dev. of 7 runs, 100 loops each) 
mask = mapped_array.values >= mapped_array.values.mean()
print(mapped_array.apply_mask(mask).values)

big_mask = big_mapped_array.values >= big_mapped_array.values.mean()
%timeit big_mapped_array.apply_mask(big_mask)
%timeit big_mapped_array_nosort.apply_mask(big_mask)
[12. 13. 14. 13. 12.] 18.2 ms ± 487 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 17.9 ms ± 143 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
@njit
def every_2_nb(inout, idxs, col, mapped_arr):
    inout[idxs[::2]] = True

print(mapped_array.map_to_mask(every_2_nb))
%timeit big_mapped_array.map_to_mask(every_2_nb)
%timeit big_mapped_array_nosort.map_to_mask(every_2_nb)
[ True False True True False True True False True] 1.11 ms ± 3.83 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 1.64 ms ± 43.2 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(mapped_array.values)

print(mapped_array.top_n_mask(1))
%timeit big_mapped_array.top_n_mask(100)

print(mapped_array.bottom_n_mask(1))
%timeit big_mapped_array.bottom_n_mask(100)
[10. 11. 12. 13. 14. 13. 12. 11. 10.] [False False True False True False True False False] 50.8 ms ± 76.3 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) [ True False False True False False False False True] 51.6 ms ± 358 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) 
print(mapped_array.top_n(1).values)
%timeit big_mapped_array.top_n(100)

print(mapped_array.bottom_n(1).values)
%timeit big_mapped_array.bottom_n(100)
[12. 14. 12.] 61.2 ms ± 642 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) [10. 13. 10.] 59.6 ms ± 362 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) 
print(mapped_array.is_expandable())

%timeit big_mapped_array.is_expandable()
True 2.44 ms ± 11.7 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd())
print(mapped_array.to_pd(fill_value=0.))

%timeit big_mapped_array.to_pd()
lvl1 0 1 lvl2 a b c d x 10.0 13.0 12.0 NaN y 11.0 14.0 11.0 NaN z 12.0 13.0 10.0 NaN lvl1 0 1 lvl2 a b c d x 10.0 13.0 12.0 0.0 y 11.0 14.0 11.0 0.0 z 12.0 13.0 10.0 0.0 5.41 ms ± 239 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array[(0, 'a')].to_pd(ignore_index=True))
%timeit big_mapped_array[0].to_pd(ignore_index=True)

print(mapped_array.to_pd(ignore_index=True))
%timeit big_mapped_array.to_pd(ignore_index=True)

print(mapped_array_grouped.to_pd(ignore_index=True))
print(mapped_array_grouped.to_pd(ignore_index=True, fill_value=0))
%timeit big_mapped_array_grouped.to_pd(ignore_index=True)
0 10.0 1 11.0 2 12.0 Name: (0, a), dtype: float64 8.29 ms ± 1.09 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) lvl1 0 1 lvl2 a b c d 0 10.0 13.0 12.0 NaN 1 11.0 14.0 11.0 NaN 2 12.0 13.0 10.0 NaN 3.48 ms ± 202 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 1 0 10.0 13.0 1 11.0 14.0 2 12.0 13.0 3 NaN 12.0 4 NaN 11.0 5 NaN 10.0 lvl1 0 1 0 10.0 13.0 1 11.0 14.0 2 12.0 13.0 3 0.0 12.0 4 0.0 11.0 5 0.0 10.0 3.34 ms ± 79.2 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
@njit
def mean_reduce_nb(col, a):
    return np.mean(a)

print(mapped_array[(0, 'a')].reduce(mean_reduce_nb))
print(mapped_array[[(0, 'a'), (1, 'b')]].reduce(mean_reduce_nb))
print(mapped_array.reduce(mean_reduce_nb))
print(mapped_array.reduce(mean_reduce_nb, fill_value=0.))
print(mapped_array.reduce(mean_reduce_nb, fill_value=0., wrap_kwargs=dict(dtype=np.int64)))
print(mapped_array.reduce(mean_reduce_nb, wrap_kwargs=dict(to_timedelta=True)))

%timeit big_mapped_array.reduce(mean_reduce_nb)
%timeit big_mapped_array_nosort.reduce(mean_reduce_nb)
11.0 lvl1 lvl2 0 a 11.000000 1 b 13.333333 Name: reduce, dtype: float64 lvl1 lvl2 0 a 11.000000 1 b 13.333333 c 11.000000 d NaN Name: reduce, dtype: float64 lvl1 lvl2 0 a 11.000000 1 b 13.333333 c 11.000000 d 0.000000 Name: reduce, dtype: float64 lvl1 lvl2 0 a 11.000000 1 b 13.333333 c 11.000000 d 0.000000 Name: reduce, dtype: float64 lvl1 lvl2 0 a 11 days 00:00:00 1 b 13 days 08:00:00 c 11 days 00:00:00 d NaT Name: reduce, dtype: timedelta64[ns] 2.07 ms ± 18.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 2.52 ms ± 38 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array_grouped[0].reduce(mean_reduce_nb))
print(mapped_array_grouped[[0]].reduce(mean_reduce_nb))
print(mapped_array_grouped.reduce(mean_reduce_nb))
print(mapped_array_grouped.reduce(mean_reduce_nb, group_by=False))

%timeit big_mapped_array_grouped.reduce(mean_reduce_nb)
11.0 lvl1 0 11.0 Name: reduce, dtype: float64 lvl1 0 11.000000 1 12.166667 Name: reduce, dtype: float64 lvl1 lvl2 0 a 11.000000 1 b 13.333333 c 11.000000 d NaN Name: reduce, dtype: float64 2.13 ms ± 141 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
@njit
def argmin_reduce_nb(col, a):
    return np.argmin(a)

print(mapped_array.reduce(argmin_reduce_nb, returns_idx=True))
%timeit big_mapped_array.reduce(argmin_reduce_nb, returns_idx=True)

print(mapped_array.reduce(argmin_reduce_nb, returns_idx=True, to_index=False))
%timeit big_mapped_array.reduce(argmin_reduce_nb, returns_idx=True, to_index=False)

print(mapped_array_grouped.reduce(argmin_reduce_nb, returns_idx=True))
%timeit big_mapped_array_grouped.reduce(argmin_reduce_nb, returns_idx=True)
lvl1 lvl2 0 a x 1 b x c z d NaN Name: reduce, dtype: object 3.9 ms ± 76.7 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 0 1 b 0 c 2 d -1 Name: reduce, dtype: int64 2.89 ms ± 113 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 x 1 z Name: reduce, dtype: object 3.15 ms ± 100 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
@njit
def min_max_reduce_nb(col, a):
    return np.array([np.min(a), np.max(a)])

print(mapped_array[(0, 'a')].reduce(min_max_reduce_nb, returns_array=True))
print(mapped_array[[(0, 'a'), (1, 'b')]].reduce(min_max_reduce_nb, returns_array=True))
print(mapped_array.reduce(min_max_reduce_nb, returns_array=True))
print(mapped_array.reduce(min_max_reduce_nb, returns_array=True, wrap_kwargs=dict(name_or_index=['min', 'max'])))
print(mapped_array.reduce(min_max_reduce_nb, returns_array=True, wrap_kwargs=dict(name_or_index=['min', 'max']), fill_value=0.))
print(mapped_array.reduce(min_max_reduce_nb, returns_array=True, wrap_kwargs=dict(to_timedelta=True)))

%timeit big_mapped_array.reduce(min_max_reduce_nb, returns_array=True)
0 10.0 1 12.0 Name: (0, a), dtype: float64 lvl1 0 1 lvl2 a b 0 10.0 13.0 1 12.0 14.0 lvl1 0 1 lvl2 a b c d 0 10.0 13.0 10.0 NaN 1 12.0 14.0 12.0 NaN lvl1 0 1 lvl2 a b c d min 10.0 13.0 10.0 NaN max 12.0 14.0 12.0 NaN lvl1 0 1 lvl2 a b c d min 10.0 13.0 10.0 0.0 max 12.0 14.0 12.0 0.0 lvl1 0 1 lvl2 a b c d 0 10 days 13 days 10 days NaT 1 12 days 14 days 12 days NaT 2.45 ms ± 12.1 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array_grouped[0].reduce(min_max_reduce_nb, returns_array=True))
print(mapped_array_grouped[[0]].reduce(min_max_reduce_nb, returns_array=True))
print(mapped_array_grouped.reduce(min_max_reduce_nb, returns_array=True))
print(mapped_array_grouped.reduce(min_max_reduce_nb, returns_array=True, group_by=False))

%timeit big_mapped_array_grouped.reduce(min_max_reduce_nb, returns_array=True)
0 10.0 1 12.0 dtype: float64 lvl1 0 0 10.0 1 12.0 lvl1 0 1 0 10.0 10.0 1 12.0 14.0 lvl1 0 1 lvl2 a b c d 0 10.0 13.0 10.0 NaN 1 12.0 14.0 12.0 NaN 2.38 ms ± 3.66 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
@njit
def idxmin_idxmax_reduce_nb(col, a):
    return np.array([np.argmin(a), np.argmax(a)])

print(mapped_array.reduce(idxmin_idxmax_reduce_nb, returns_array=True, returns_idx=True))
%timeit big_mapped_array.reduce(idxmin_idxmax_reduce_nb, returns_array=True, returns_idx=True)

print(mapped_array.reduce(idxmin_idxmax_reduce_nb, returns_array=True, returns_idx=True, to_index=False))
%timeit big_mapped_array.reduce(idxmin_idxmax_reduce_nb, returns_array=True, returns_idx=True, to_index=False)

print(mapped_array_grouped.reduce(idxmin_idxmax_reduce_nb, returns_array=True, returns_idx=True))
%timeit big_mapped_array_grouped.reduce(idxmin_idxmax_reduce_nb, returns_array=True, returns_idx=True)
lvl1 0 1 lvl2 a b c d 0 x x z NaN 1 z y x NaN 82.4 ms ± 388 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) lvl1 0 1 lvl2 a b c d 0 0 0 2 -1 1 2 1 0 -1 4.04 ms ± 17.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 1 0 x z 1 z y 4.45 ms ± 3.78 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.nth(0))
print(mapped_array.nth(-1))
%timeit big_mapped_array.nth(0)

print(mapped_array_grouped.nth(0))
%timeit big_mapped_array_grouped.nth(0)
lvl1 lvl2 0 a 10.0 1 b 13.0 c 12.0 d NaN Name: nth, dtype: float64 lvl1 lvl2 0 a 12.0 1 b 13.0 c 10.0 d NaN Name: nth, dtype: float64 1.03 ms ± 1.44 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) lvl1 0 10.0 1 13.0 Name: nth, dtype: float64 1.11 ms ± 3.55 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(mapped_array.to_pd().vbt.min())
%timeit big_mapped_array.to_pd().vbt.min()

print(mapped_array.min())
%timeit big_mapped_array.min()

print(mapped_array_grouped.min())
%timeit big_mapped_array_grouped.min()
lvl1 lvl2 0 a 10.0 1 b 13.0 c 10.0 d NaN Name: min, dtype: float64 10.1 ms ± 458 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 10.0 1 b 13.0 c 10.0 d NaN Name: min, dtype: float64 2.98 ms ± 4.29 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 10.0 1 10.0 Name: min, dtype: float64 3.02 ms ± 5.68 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.max())
%timeit big_mapped_array.to_pd().vbt.max()

print(mapped_array.max())
%timeit big_mapped_array.max()

print(mapped_array_grouped.max())
%timeit big_mapped_array_grouped.max()
lvl1 lvl2 0 a 12.0 1 b 14.0 c 12.0 d NaN Name: max, dtype: float64 9.97 ms ± 92 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 12.0 1 b 14.0 c 12.0 d NaN Name: max, dtype: float64 2.98 ms ± 6.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 12.0 1 14.0 Name: max, dtype: float64 3.03 ms ± 3.58 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.mean())
%timeit big_mapped_array.to_pd().vbt.mean()

print(mapped_array.mean())
%timeit big_mapped_array.mean()

print(mapped_array_grouped.mean())
%timeit big_mapped_array_grouped.mean()
lvl1 lvl2 0 a 11.000000 1 b 13.333333 c 11.000000 d NaN Name: mean, dtype: float64 8.4 ms ± 336 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 11.000000 1 b 13.333333 c 11.000000 d NaN Name: mean, dtype: float64 3.27 ms ± 10.9 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 11.000000 1 12.166667 Name: mean, dtype: float64 3.32 ms ± 5.36 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.median())
%timeit big_mapped_array.to_pd().vbt.median()

print(mapped_array.median())
%timeit big_mapped_array.median()

print(mapped_array_grouped.median())
%timeit big_mapped_array_grouped.median()
lvl1 lvl2 0 a 11.0 1 b 13.0 c 11.0 d NaN Name: median, dtype: float64 16.9 ms ± 151 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 11.0 1 b 13.0 c 11.0 d NaN Name: median, dtype: float64 12.5 ms ± 7.77 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 11.0 1 12.5 Name: median, dtype: float64 11.7 ms ± 13.3 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.std())
print(mapped_array.to_pd().vbt.std(ddof=0))
%timeit big_mapped_array.to_pd().vbt.std()

print(mapped_array.std())
print(mapped_array.std(ddof=0))
%timeit big_mapped_array.std()

print(mapped_array_grouped.std())
%timeit big_mapped_array_grouped.std()
lvl1 lvl2 0 a 1.00000 1 b 0.57735 c 1.00000 d NaN Name: std, dtype: float64 lvl1 lvl2 0 a 0.816497 1 b 0.471405 c 0.816497 d NaN Name: std, dtype: float64 8.7 ms ± 23.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 1.00000 1 b 0.57735 c 1.00000 d NaN Name: std, dtype: float64 lvl1 lvl2 0 a 0.816497 1 b 0.471405 c 0.816497 d NaN Name: std, dtype: float64 6.59 ms ± 134 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 1.00000 1 1.47196 Name: std, dtype: float64 6.24 ms ± 3.6 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.sum())
%timeit big_mapped_array.to_pd().vbt.sum()

print(mapped_array.sum())
%timeit big_mapped_array.sum()

print(mapped_array_grouped.sum())
%timeit big_mapped_array_grouped.sum()
lvl1 lvl2 0 a 33.0 1 b 40.0 c 33.0 d 0.0 Name: sum, dtype: float64 8.63 ms ± 36.9 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 33.0 1 b 40.0 c 33.0 d 0.0 Name: sum, dtype: float64 3.26 ms ± 6.28 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 33.0 1 73.0 Name: sum, dtype: float64 3.31 ms ± 3.33 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.idxmin())
%timeit big_mapped_array.to_pd().vbt.idxmin()

print(mapped_array.idxmin())
%timeit big_mapped_array.idxmin()

print(mapped_array_grouped.idxmin())
%timeit big_mapped_array_grouped.idxmin()
lvl1 lvl2 0 a x 1 b x c z d NaN Name: idxmin, dtype: object 9.22 ms ± 130 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a x 1 b x c z d NaN Name: idxmin, dtype: object 4.86 ms ± 28.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 x 1 z Name: idxmin, dtype: object 4.21 ms ± 7.54 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.idxmax())
%timeit big_mapped_array.to_pd().vbt.idxmax()

print(mapped_array.idxmax())
%timeit big_mapped_array.idxmax()

print(mapped_array_grouped.idxmax())
%timeit big_mapped_array_grouped.idxmax()
lvl1 lvl2 0 a z 1 b y c x d NaN Name: idxmax, dtype: object 9.4 ms ± 395 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a z 1 b y c x d NaN Name: idxmax, dtype: object 4.88 ms ± 41.7 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 z 1 y Name: idxmax, dtype: object 4.22 ms ± 7.03 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array.to_pd().vbt.describe())
print(mapped_array.to_pd().vbt.describe(percentiles=[0.3, 0.7]))
%timeit big_mapped_array.to_pd().vbt.describe()

print(mapped_array.describe())
print(mapped_array.describe(percentiles=[0.3, 0.7]))
%timeit big_mapped_array.describe()

print(mapped_array_grouped.describe())
%timeit big_mapped_array_grouped.describe()
lvl1 0 1 lvl2 a b c d count 3.0 3.000000 3.0 0.0 mean 11.0 13.333333 11.0 NaN std 1.0 0.577350 1.0 NaN min 10.0 13.000000 10.0 NaN 25% 10.5 13.000000 10.5 NaN 50% 11.0 13.000000 11.0 NaN 75% 11.5 13.500000 11.5 NaN max 12.0 14.000000 12.0 NaN lvl1 0 1 lvl2 a b c d count 3.0 3.000000 3.0 0.0 mean 11.0 13.333333 11.0 NaN std 1.0 0.577350 1.0 NaN min 10.0 13.000000 10.0 NaN 30% 10.6 13.000000 10.6 NaN 50% 11.0 13.000000 11.0 NaN 70% 11.4 13.400000 11.4 NaN max 12.0 14.000000 12.0 NaN 41.4 ms ± 24.4 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) lvl1 0 1 lvl2 a b c d count 3.0 3.000000 3.0 0.0 mean 11.0 13.333333 11.0 NaN std 1.0 0.577350 1.0 NaN min 10.0 13.000000 10.0 NaN 25% 10.5 13.000000 10.5 NaN 50% 11.0 13.000000 11.0 NaN 75% 11.5 13.500000 11.5 NaN max 12.0 14.000000 12.0 NaN lvl1 0 1 lvl2 a b c d count 3.0 3.000000 3.0 0.0 mean 11.0 13.333333 11.0 NaN std 1.0 0.577350 1.0 NaN min 10.0 13.000000 10.0 NaN 30% 10.6 13.000000 10.6 NaN 50% 11.0 13.000000 11.0 NaN 70% 11.4 13.400000 11.4 NaN max 12.0 14.000000 12.0 NaN 35.6 ms ± 432 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) lvl1 0 1 count 3.0 6.000000 mean 11.0 12.166667 std 1.0 1.471960 min 10.0 10.000000 25% 10.5 11.250000 50% 11.0 12.500000 75% 11.5 13.000000 max 12.0 14.000000 37.9 ms ± 123 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) 
print(mapped_array.to_pd().vbt.count())
%timeit big_mapped_array.to_pd().vbt.count()

print(mapped_array.count())
%timeit big_mapped_array.count()

print(mapped_array_grouped.count())
%timeit big_mapped_array_grouped.count()
lvl1 lvl2 0 a 3 1 b 3 c 3 d 0 Name: count, dtype: int64 7.45 ms ± 215 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 lvl2 0 a 3 1 b 3 c 3 d 0 Name: count, dtype: int64 49.6 µs ± 200 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each) lvl1 0 3 1 6 Name: count, dtype: int64 48.3 µs ± 83.7 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each) 
mapping = {x: str(x) for x in np.unique(mapped_array.values)}
big_mapping = {x: str(x) for x in np.unique(big_mapped_array.values)}

print(mapped_array[(0, 'a')].value_counts())
%timeit big_mapped_array[0].value_counts()

print(mapped_array[(0, 'a')].value_counts(mapping=mapping))
%timeit big_mapped_array[0].value_counts(mapping=big_mapping)

print(mapped_array.value_counts())
%timeit big_mapped_array.value_counts()

print(mapped_array.value_counts(mapping=mapping))
%timeit big_mapped_array.value_counts(mapping=big_mapping)

print(mapped_array_grouped.value_counts())
%timeit big_mapped_array_grouped.value_counts()
10.0 1 11.0 1 12.0 1 Name: (0, a), dtype: int64 13.4 ms ± 584 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 10.0 1 11.0 1 12.0 1 Name: (0, a), dtype: int64 12.9 ms ± 353 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 1 lvl2 a b c d 10.0 1 0 1 0 11.0 1 0 1 0 12.0 1 0 1 0 13.0 0 2 0 0 14.0 0 1 0 0 10.9 ms ± 23.5 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 1 lvl2 a b c d 10.0 1 0 1 0 11.0 1 0 1 0 12.0 1 0 1 0 13.0 0 2 0 0 14.0 0 1 0 0 11.2 ms ± 121 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) lvl1 0 1 10.0 1 1 11.0 1 1 12.0 1 1 13.0 0 2 14.0 0 1 10.9 ms ± 451 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(mapped_array[(0, 'a')].stats())
%timeit big_mapped_array[0].stats(silence_warnings=True)

print(mapped_array.stats(column=(0, 'a')))
%timeit big_mapped_array.stats(column=0, silence_warnings=True)

print(mapped_array.stats())
%timeit big_mapped_array.stats(silence_warnings=True)

print(mapped_array.copy(mapping=mapping).stats())
%timeit big_mapped_array.copy(mapping=big_mapping).stats(silence_warnings=True)
Start x End z Period 3 days 00:00:00 Count 3 Mean 11.0 Std 1.0 Min 10.0 Median 11.0 Max 12.0 Min Index x Max Index z Name: (0, a), dtype: object 29 ms ± 287 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) Start x End z Period 3 days 00:00:00 Count 3 Mean 11.0 Std 1.0 Min 10.0 Median 11.0 Max 12.0 Min Index x Max Index z Name: (0, a), dtype: object 46 ms ± 1.01 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) Start x End z Period 3 days 00:00:00 Count 2.25 Mean 11.777778 Std 0.859117 Min 11.0 Median 11.666667 Max 12.666667 Name: agg_func_mean, dtype: object
/Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Metric 'idx_min' returned multiple values despite having no aggregation function import sys /Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Metric 'idx_max' returned multiple values despite having no aggregation function import sys /Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Object has multiple columns. Aggregating using <function mean at 0x7feb1013b598>. Pass column to select a single column/group. import sys
44.3 ms ± 343 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) Start x End z Period 3 days 00:00:00 Count 2.25 Value Counts: 10.0 0.5 Value Counts: 11.0 0.5 Value Counts: 12.0 0.5 Value Counts: 13.0 0.5 Value Counts: 14.0 0.25 Name: agg_func_mean, dtype: object
/Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:10: UserWarning: Object has multiple columns. Aggregating using <function mean at 0x7feb1013b598>. Pass column to select a single column/group. # Remove the CWD from sys.path while we load stuff.
23.1 ms ± 475 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) 
mapped_array[(0, 'a')].histplot().show_svg()
mapped_array.histplot().show_svg()
mapped_array_grouped.histplot().show_svg()
Image in a Jupyter notebookImage in a Jupyter notebookImage in a Jupyter notebook
mapped_array[(0, 'a')].boxplot().show_svg()
mapped_array.boxplot().show_svg()
mapped_array_grouped.boxplot().show_svg()
Image in a Jupyter notebookImage in a Jupyter notebookImage in a Jupyter notebook
mapped_array[(0, 'a')].plots().show_svg()
mapped_array.plots().show_svg()
Image in a Jupyter notebookImage in a Jupyter notebook

Records

print(records[(0, 'a')].values)
print(records[(0, 'a')].wrapper.columns)

print(records[(1, 'b')].values)
print(records[(1, 'b')].wrapper.columns)

print(records[[(0, 'a'), (0, 'a')]].values)
print(records[[(0, 'a'), (0, 'a')]].wrapper.columns)

print(records[[(0, 'a'), (1, 'b')]].values)
print(records[[(0, 'a'), (1, 'b')]].wrapper.columns)

%timeit big_records.iloc[0]
%timeit big_records.iloc[:]
[(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.)] MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) [(3, 0, 0, 13., 18.) (4, 1, 0, 14., 17.) (5, 2, 0, 13., 18.)] MultiIndex([(1, 'b')], names=['lvl1', 'lvl2']) [(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (0, 0, 1, 10., 21.) (1, 1, 1, 11., 20.) (2, 2, 1, 12., 19.)] MultiIndex([(0, 'a'), (0, 'a')], names=['lvl1', 'lvl2']) [(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.)] MultiIndex([(0, 'a'), (1, 'b')], names=['lvl1', 'lvl2']) 810 µs ± 9.11 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 13.7 ms ± 259 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(records_nosort[(0, 'a')].values)
print(records_nosort[(0, 'a')].wrapper.columns)

print(records_nosort[(1, 'b')].values)
print(records_nosort[(1, 'b')].wrapper.columns)

print(records_nosort[[(0, 'a'), (0, 'a')]].values)
print(records_nosort[[(0, 'a'), (0, 'a')]].wrapper.columns)

print(records_nosort[[(0, 'a'), (1, 'b')]].values)
print(records_nosort[[(0, 'a'), (1, 'b')]].wrapper.columns)

%timeit big_records_nosort.iloc[0]
%timeit big_records_nosort.iloc[:]
[(2, 2, 0, 12., 19.) (1, 1, 0, 11., 20.) (0, 0, 0, 10., 21.)] MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) [(5, 2, 0, 13., 18.) (4, 1, 0, 14., 17.) (3, 0, 0, 13., 18.)] MultiIndex([(1, 'b')], names=['lvl1', 'lvl2']) [(2, 2, 0, 12., 19.) (1, 1, 0, 11., 20.) (0, 0, 0, 10., 21.) (2, 2, 1, 12., 19.) (1, 1, 1, 11., 20.) (0, 0, 1, 10., 21.)] MultiIndex([(0, 'a'), (0, 'a')], names=['lvl1', 'lvl2']) [(2, 2, 0, 12., 19.) (1, 1, 0, 11., 20.) (0, 0, 0, 10., 21.) (5, 2, 1, 13., 18.) (4, 1, 1, 14., 17.) (3, 0, 1, 13., 18.)] MultiIndex([(0, 'a'), (1, 'b')], names=['lvl1', 'lvl2']) 842 µs ± 5.42 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 15.9 ms ± 678 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(records_grouped[0].wrapper.columns)  # indexing on groups, not columns!
print(records_grouped[0].wrapper.ndim)
print(records_grouped[0].wrapper.grouped_ndim)
print(records_grouped[0].wrapper.grouper.group_by)

print(records_grouped[1].wrapper.columns)
print(records_grouped[1].wrapper.ndim)
print(records_grouped[1].wrapper.grouped_ndim)
print(records_grouped[1].wrapper.grouper.group_by)

print(records_grouped[[0]].wrapper.columns)
print(records_grouped[[0]].wrapper.ndim)
print(records_grouped[[0]].wrapper.grouped_ndim)
print(records_grouped[[0]].wrapper.grouper.group_by)

print(records_grouped[[0, 1]].wrapper.columns)
print(records_grouped[[0, 1]].wrapper.ndim)
print(records_grouped[[0, 1]].wrapper.grouped_ndim)
print(records_grouped[[0, 1]].wrapper.grouper.group_by)

%timeit big_records_grouped.iloc[0]
%timeit big_records_grouped.iloc[:]
MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) 1 1 Index([0], dtype='int64', name='lvl1') MultiIndex([(1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 1 Index([1, 1, 1], dtype='int64', name='lvl1') MultiIndex([(0, 'a')], names=['lvl1', 'lvl2']) 2 2 Index([0], dtype='int64', name='lvl1') MultiIndex([(0, 'a'), (1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 2 Index([0, 1, 1, 1], dtype='int64', name='lvl1') 6.25 ms ± 70.1 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 13.9 ms ± 264 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(records.wrapper.index)
print(records.wrapper.columns)
print(records.wrapper.ndim)
print(records.wrapper.grouper.group_by)

print(records_grouped.wrapper.index)
print(records_grouped.wrapper.columns)
print(records_grouped.wrapper.ndim)
print(records_grouped.wrapper.grouper.group_by)
Index(['x', 'y', 'z'], dtype='object') MultiIndex([(0, 'a'), (1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 None Index(['x', 'y', 'z'], dtype='object') MultiIndex([(0, 'a'), (1, 'b'), (1, 'c'), (1, 'd')], names=['lvl1', 'lvl2']) 2 Index([0, 1, 1, 1], dtype='int64', name='lvl1') 
print(records.values)

print(records.recarray)
%timeit big_records.recarray
[(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.) (7, 1, 2, 11., 20.) (8, 2, 2, 10., 21.)] [(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.) (7, 1, 2, 11., 20.) (8, 2, 2, 10., 21.)] 3.54 µs ± 13.6 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.records)
id idx col some_field1 some_field2 0 0 0 0 10.0 21.0 1 1 1 0 11.0 20.0 2 2 2 0 12.0 19.0 3 3 0 1 13.0 18.0 4 4 1 1 14.0 17.0 5 5 2 1 13.0 18.0 6 6 0 2 12.0 19.0 7 7 1 2 11.0 20.0 8 8 2 2 10.0 21.0 
print(records.is_sorted())
%timeit big_records.is_sorted()

print(records_nosort.is_sorted())
%timeit big_records_nosort.is_sorted()

print(records.is_sorted(incl_id=True))
%timeit big_records.is_sorted(incl_id=True)

print(records_nosort.is_sorted(incl_id=True))
%timeit big_records_nosort.is_sorted(incl_id=True)
True 1.06 ms ± 1.99 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) False 8.89 µs ± 232 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) True 1.56 ms ± 1.74 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) False 8.41 µs ± 37.3 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each) 
print(records.sort().records_arr)
%timeit big_records.sort()

print(records_nosort.sort().records_arr)
%timeit big_records_nosort.sort()

print(records.sort(incl_id=True).records_arr)
%timeit big_records.sort(incl_id=True)

print(records_nosort.sort(incl_id=True).records_arr)
%timeit big_records_nosort.sort(incl_id=True)
[(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.) (7, 1, 2, 11., 20.) (8, 2, 2, 10., 21.)] 1.3 ms ± 771 ns per loop (mean ± std. dev. of 7 runs, 1000 loops each) [(2, 2, 0, 12., 19.) (1, 1, 0, 11., 20.) (0, 0, 0, 10., 21.) (5, 2, 1, 13., 18.) (4, 1, 1, 14., 17.) (3, 0, 1, 13., 18.) (8, 2, 2, 10., 21.) (7, 1, 2, 11., 20.) (6, 0, 2, 12., 19.)] 70.6 ms ± 1.59 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) [(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.) (7, 1, 2, 11., 20.) (8, 2, 2, 10., 21.)] 1.83 ms ± 12.4 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) [(0, 0, 0, 10., 21.) (1, 1, 0, 11., 20.) (2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.) (7, 1, 2, 11., 20.) (8, 2, 2, 10., 21.)] 66 ms ± 1.53 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) 
mask = records.values['some_field1'] >= records.values['some_field1'].mean()
print(records.apply_mask(mask).values)

big_mask = big_records.values['some_field1'] >= big_records.values['some_field1'].mean()
%timeit big_records.apply_mask(big_mask)
[(2, 2, 0, 12., 19.) (3, 0, 1, 13., 18.) (4, 1, 1, 14., 17.) (5, 2, 1, 13., 18.) (6, 0, 2, 12., 19.)] 10.4 ms ± 232 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
@njit
def map_nb(record):
    return record.some_field1 + record.some_field2 * 2

print(records.map(map_nb).sum())
print(records_grouped.map(map_nb).sum())
print(records_grouped.map(map_nb, group_by=False).sum())

%timeit vbt.records.MappedArray(\
    big_wrapper,\
    big_records_arr['some_field1'] + big_records_arr['some_field2'] * 2,\
    big_records_arr['col'],\
)
%timeit big_records.map(map_nb)  # faster
lvl1 lvl2 0 a 153.0 1 b 146.0 c 153.0 d 0.0 Name: sum, dtype: float64 lvl1 0 153.0 1 299.0 Name: sum, dtype: float64 lvl1 lvl2 0 a 153.0 1 b 146.0 c 153.0 d 0.0 Name: sum, dtype: float64 3.25 ms ± 735 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 4.53 ms ± 278 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(records.map_field('col').values)
print(records.map_field('idx').values)
print(records.map_field('some_field1').values)
print(records.map_field('some_field2').values)

print(records.map_field('some_field1').sum())
print(records_grouped.map_field('some_field1').sum())
print(records_grouped.map_field('some_field1', group_by=False).sum())

%timeit big_records.map_field('some_field1')
[0 0 0 1 1 1 2 2 2] [0 1 2 0 1 2 0 1 2] [10. 11. 12. 13. 14. 13. 12. 11. 10.] [21. 20. 19. 18. 17. 18. 19. 20. 21.] lvl1 lvl2 0 a 33.0 1 b 40.0 c 33.0 d 0.0 Name: sum, dtype: float64 lvl1 0 33.0 1 73.0 Name: sum, dtype: float64 lvl1 lvl2 0 a 33.0 1 b 40.0 c 33.0 d 0.0 Name: sum, dtype: float64 293 µs ± 4.43 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(records.map_array(records_arr['some_field1'] + records_arr['some_field2'] * 2).values)
print(records_grouped.map_array(records_arr['some_field1'] + records_arr['some_field2'] * 2).sum())
print(records_grouped.map_array(records_arr['some_field1'] + records_arr['some_field2'] * 2, group_by=False).sum())

%timeit big_records.map_array(big_records_arr['some_field1'] + big_records_arr['some_field2'] * 2)
[52. 51. 50. 49. 48. 49. 50. 51. 52.] lvl1 0 153.0 1 299.0 Name: sum, dtype: float64 lvl1 lvl2 0 a 153.0 1 b 146.0 c 153.0 d 0.0 Name: sum, dtype: float64 2.55 ms ± 544 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(records.count())
print(records_grouped.count())
print(records_grouped.count(group_by=False))

%timeit big_records.count()
%timeit big_records_grouped.count()
lvl1 lvl2 0 a 3 1 b 3 c 3 d 0 Name: count, dtype: int64 lvl1 0 3 1 6 Name: count, dtype: int64 lvl1 lvl2 0 a 3 1 b 3 c 3 d 0 Name: count, dtype: int64 49.5 µs ± 133 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each) 48.5 µs ± 595 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each) 
filter_mask = np.array([True, False, False, False, False, False, False, False, True])

print(records.apply_mask(filter_mask).count())
print(records_grouped.apply_mask(filter_mask).count())
print(records_grouped.apply_mask(filter_mask, group_by=False).count())
lvl1 lvl2 0 a 1 1 b 0 c 1 d 0 Name: count, dtype: int64 lvl1 0 1 1 1 Name: count, dtype: int64 lvl1 lvl2 0 a 1 1 b 0 c 1 d 0 Name: count, dtype: int64 
filtered_records = records.apply_mask(filter_mask)
print(filtered_records.records)
id idx col some_field1 some_field2 0 0 0 0 10.0 21.0 1 8 2 2 10.0 21.0 
print(filtered_records[(0, 'a')].records)
print(filtered_records[(0, 'a')].map_field('some_field1').values)
print(filtered_records[(0, 'a')].map_field('some_field1').min())
print(filtered_records[(0, 'a')].count())
id idx col some_field1 some_field2 0 0 0 0 10.0 21.0 [10.] 10.0 1 
print(filtered_records[(1, 'b')].records)
print(filtered_records[(1, 'b')].map_field('some_field1').values)
print(filtered_records[(1, 'b')].map_field('some_field1').min())
print(filtered_records[(1, 'b')].count())
Empty DataFrame Columns: [id, idx, col, some_field1, some_field2] Index: [] [] nan 0 
print(filtered_records[(1, 'c')].records)
print(filtered_records[(1, 'c')].map_field('some_field1').values)
print(filtered_records[(1, 'c')].map_field('some_field1').min())
print(filtered_records[(1, 'c')].count())
id idx col some_field1 some_field2 0 8 2 0 10.0 21.0 [10.] 10.0 1 
print(filtered_records[(1, 'd')].records)
print(filtered_records[(1, 'd')].map_field('some_field1').values)
print(filtered_records[(1, 'd')].map_field('some_field1').min())
print(filtered_records[(1, 'd')].count())
Empty DataFrame Columns: [id, idx, col, some_field1, some_field2] Index: [] [] nan 0 
print(records[(0, 'a')].stats())
%timeit big_records[0].stats(silence_warnings=True)

print(records.stats(column=(0, 'a')))
%timeit big_records.stats(column=0, silence_warnings=True)

print(records.stats())
%timeit big_records.stats(silence_warnings=True)
Start x End z Period 3 days 00:00:00 Count 3 Name: (0, a), dtype: object 12.7 ms ± 680 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) Start x End z Period 3 days 00:00:00 Count 3 Name: (0, a), dtype: object 2.18 ms ± 42.2 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) Start x End z Period 3 days 00:00:00 Count 2.25 Name: agg_func_mean, dtype: object 2.1 ms ± 1.67 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
/Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Object has multiple columns. Aggregating using <function mean at 0x7fbd30c34598>. Pass column to select a single column/group. 
records[(0, 'a')].plots()
/Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: UserWarning: No subplots to plot

Drawdowns

ts = pd.DataFrame({
    'a': [2, 1, 3, 1, 4, 1], 
    'b': [1, 2, 1, 3, 1, 4],
    'c': [1, 2, 3, 2, 1, 2],
    'd': [1, 2, 3, 4, 5, 6]
}, index=[
    datetime(2020, 1, 1),
    datetime(2020, 1, 2),
    datetime(2020, 1, 3),
    datetime(2020, 1, 4),
    datetime(2020, 1, 5),
    datetime(2020, 1, 6)
])

np.random.seed(42)
big_ts = pd.DataFrame(np.random.randint(1, 10, size=(1000, 1000)))

drawdowns = vbt.Drawdowns.from_ts(ts, freq='1 days')
print(drawdowns.values.shape)
(6,) 
big_drawdowns = vbt.Drawdowns.from_ts(big_ts, freq='1 days')
print(big_drawdowns.values.shape)
(99845,) 
%timeit vbt.Drawdowns.from_ts(big_ts, freq='1 days')
5.37 ms ± 57.5 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns.records)
print(drawdowns.ts)

print(drawdowns['a'].records)
print(drawdowns['a'].ts)

%timeit big_drawdowns.iloc[0]
%timeit big_drawdowns.iloc[:]
id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 0 0 0 1 1 2 2.0 1.0 1 1 0 2 3 3 4 3.0 1.0 2 2 0 4 5 5 5 4.0 1.0 3 3 1 1 2 2 3 2.0 1.0 4 4 1 3 4 4 5 3.0 1.0 5 5 2 2 3 4 5 3.0 1.0 end_val status 0 2.0 1 1 3.0 1 2 1.0 0 3 2.0 1 4 3.0 1 5 2.0 0 a b c d 2020-01-01 2 1 1 1 2020-01-02 1 2 2 2 2020-01-03 3 1 3 3 2020-01-04 1 3 2 4 2020-01-05 4 1 1 5 2020-01-06 1 4 2 6 id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 0 0 0 1 1 2 2.0 1.0 1 1 0 2 3 3 4 3.0 1.0 2 2 0 4 5 5 5 4.0 1.0 end_val status 0 2.0 1 1 3.0 1 2 1.0 0 2020-01-01 2 2020-01-02 1 2020-01-03 3 2020-01-04 1 2020-01-05 4 2020-01-06 1 Name: a, dtype: int64 1.21 ms ± 12.6 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 3.81 ms ± 418 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns.records_readable)
Drawdown Id Column Peak Timestamp Start Timestamp Valley Timestamp \ 0 0 a 2020-01-01 2020-01-02 2020-01-02 1 1 a 2020-01-03 2020-01-04 2020-01-04 2 2 a 2020-01-05 2020-01-06 2020-01-06 3 3 b 2020-01-02 2020-01-03 2020-01-03 4 4 b 2020-01-04 2020-01-05 2020-01-05 5 5 c 2020-01-03 2020-01-04 2020-01-05 End Timestamp Peak Value Valley Value End Value Status 0 2020-01-03 2.0 1.0 2.0 Recovered 1 2020-01-05 3.0 1.0 3.0 Recovered 2 2020-01-06 4.0 1.0 1.0 Active 3 2020-01-04 2.0 1.0 2.0 Recovered 4 2020-01-06 3.0 1.0 3.0 Recovered 5 2020-01-06 3.0 1.0 2.0 Active 
print(drawdowns['a'].count())

print(drawdowns.count())
print(drawdowns.count(group_by=group_by))
3 a 3 b 2 c 1 d 0 Name: count, dtype: int64 group first 5 second 1 Name: count, dtype: int64 
print(drawdowns['a'].drawdown.to_pd())
print(drawdowns.drawdown.to_pd())

%timeit big_drawdowns.drawdown
2020-01-01 NaN 2020-01-02 NaN 2020-01-03 -0.500000 2020-01-04 NaN 2020-01-05 -0.666667 2020-01-06 -0.750000 Name: a, dtype: float64 a b c d 2020-01-01 NaN NaN NaN NaN 2020-01-02 NaN NaN NaN NaN 2020-01-03 -0.500000 NaN NaN NaN 2020-01-04 NaN -0.500000 NaN NaN 2020-01-05 -0.666667 NaN NaN NaN 2020-01-06 -0.750000 -0.666667 -0.666667 NaN 463 µs ± 21.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].avg_drawdown())
print(drawdowns.avg_drawdown())
%timeit big_drawdowns.avg_drawdown()

print(drawdowns.avg_drawdown(group_by=group_by))
%timeit big_drawdowns.avg_drawdown(group_by=big_group_by)
-0.6388888888888888 a -0.638889 b -0.583333 c -0.666667 d NaN Name: avg_drawdown, dtype: float64 1.12 ms ± 200 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first -0.616667 second -0.666667 Name: avg_drawdown, dtype: float64 4.65 ms ± 1.91 ms per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].max_drawdown())
print(drawdowns.max_drawdown())
%timeit big_drawdowns.max_drawdown()

print(drawdowns.max_drawdown(group_by=group_by))
%timeit big_drawdowns.max_drawdown(group_by=big_group_by)
-0.75 a -0.750000 b -0.666667 c -0.666667 d NaN Name: max_drawdown, dtype: float64 1.15 ms ± 214 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first -0.750000 second -0.666667 Name: max_drawdown, dtype: float64 3.81 ms ± 652 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].recovery_return.to_pd())
print(drawdowns.recovery_return.to_pd())

%timeit big_drawdowns.recovery_return
2020-01-01 NaN 2020-01-02 NaN 2020-01-03 1.0 2020-01-04 NaN 2020-01-05 2.0 2020-01-06 0.0 Name: a, dtype: float64 a b c d 2020-01-01 NaN NaN NaN NaN 2020-01-02 NaN NaN NaN NaN 2020-01-03 1.0 NaN NaN NaN 2020-01-04 NaN 1.0 NaN NaN 2020-01-05 2.0 NaN NaN NaN 2020-01-06 0.0 2.0 1.0 NaN 479 µs ± 45.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].avg_recovery_return())
print(drawdowns.avg_recovery_return())
%timeit big_drawdowns.avg_recovery_return()

print(drawdowns.avg_recovery_return(group_by=group_by))
%timeit big_drawdowns.avg_recovery_return(group_by=big_group_by)
1.0 a 1.0 b 1.5 c 1.0 d NaN Name: avg_recovery_return, dtype: float64 1.01 ms ± 8.88 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first 1.2 second 1.0 Name: avg_recovery_return, dtype: float64 3.48 ms ± 80 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].max_recovery_return())
print(drawdowns.max_recovery_return())
%timeit big_drawdowns.max_recovery_return()

print(drawdowns.max_recovery_return(group_by=group_by))
%timeit big_drawdowns.max_recovery_return(group_by=big_group_by)
2.0 a 2.0 b 2.0 c 1.0 d NaN Name: max_recovery_return, dtype: float64 1.06 ms ± 35.7 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first 2.0 second 1.0 Name: max_recovery_return, dtype: float64 3.72 ms ± 302 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].duration.to_pd())
print(drawdowns.duration.to_pd())

%timeit big_drawdowns.duration
2020-01-01 NaN 2020-01-02 NaN 2020-01-03 1.0 2020-01-04 NaN 2020-01-05 1.0 2020-01-06 1.0 Name: a, dtype: float64 a b c d 2020-01-01 NaN NaN NaN NaN 2020-01-02 NaN NaN NaN NaN 2020-01-03 1.0 NaN NaN NaN 2020-01-04 NaN 1.0 NaN NaN 2020-01-05 1.0 NaN NaN NaN 2020-01-06 1.0 1.0 3.0 NaN 483 µs ± 6.53 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].avg_duration())
print(drawdowns.avg_duration())
%timeit big_drawdowns.avg_duration()

print(drawdowns.avg_duration(group_by=group_by))
%timeit big_drawdowns.avg_duration(group_by=big_group_by)
1 days 00:00:00 a 1 days b 1 days c 3 days d NaT Name: avg_duration, dtype: timedelta64[ns]
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/base/array_wrapper.py:661: UserWarning: Couldn't parse the frequency of index. Pass it as `freq` or define it globally under `settings.array_wrapper`.
905 µs ± 3.32 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first 1 days second 3 days Name: avg_duration, dtype: timedelta64[ns] 3.34 ms ± 35.3 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].max_duration())
print(drawdowns.max_duration())
%timeit big_drawdowns.max_duration()

print(drawdowns.max_duration(group_by=group_by))
%timeit big_drawdowns.max_duration(group_by=big_group_by)
1 days 00:00:00 a 1 days b 1 days c 3 days d NaT Name: max_duration, dtype: timedelta64[ns]
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/base/array_wrapper.py:661: UserWarning: Couldn't parse the frequency of index. Pass it as `freq` or define it globally under `settings.array_wrapper`.
992 µs ± 72.6 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first 1 days second 3 days Name: max_duration, dtype: timedelta64[ns] 3.35 ms ± 18.6 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].coverage())
print(drawdowns.coverage())
%timeit big_drawdowns.coverage()

print(drawdowns.coverage(group_by=group_by))
%timeit big_drawdowns.coverage(group_by=big_group_by)
0.5 a 0.500000 b 0.333333 c 0.500000 d NaN Name: coverage, dtype: float64 13.8 ms ± 181 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) group first 0.416667 second 0.250000 Name: coverage, dtype: float64 16.8 ms ± 164 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) 
print(drawdowns['a'].decline_duration.to_pd())
print(drawdowns.decline_duration.to_pd())

%timeit big_drawdowns.decline_duration
2020-01-01 NaN 2020-01-02 NaN 2020-01-03 1.0 2020-01-04 NaN 2020-01-05 1.0 2020-01-06 1.0 Name: a, dtype: float64 a b c d 2020-01-01 NaN NaN NaN NaN 2020-01-02 NaN NaN NaN NaN 2020-01-03 1.0 NaN NaN NaN 2020-01-04 NaN 1.0 NaN NaN 2020-01-05 1.0 NaN NaN NaN 2020-01-06 1.0 1.0 2.0 NaN 518 µs ± 30.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].recovery_duration.to_pd())
print(drawdowns.recovery_duration.to_pd())

%timeit big_drawdowns.recovery_duration
2020-01-01 NaN 2020-01-02 NaN 2020-01-03 1.0 2020-01-04 NaN 2020-01-05 1.0 2020-01-06 0.0 Name: a, dtype: float64 a b c d 2020-01-01 NaN NaN NaN NaN 2020-01-02 NaN NaN NaN NaN 2020-01-03 1.0 NaN NaN NaN 2020-01-04 NaN 1.0 NaN NaN 2020-01-05 1.0 NaN NaN NaN 2020-01-06 0.0 1.0 1.0 NaN 479 µs ± 17.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].recovery_duration_ratio.to_pd())
print(drawdowns.recovery_duration_ratio.to_pd())

%timeit big_drawdowns.recovery_duration_ratio
2020-01-01 NaN 2020-01-02 NaN 2020-01-03 1.0 2020-01-04 NaN 2020-01-05 1.0 2020-01-06 0.0 Name: a, dtype: float64 a b c d 2020-01-01 NaN NaN NaN NaN 2020-01-02 NaN NaN NaN NaN 2020-01-03 1.0 NaN NaN NaN 2020-01-04 NaN 1.0 NaN NaN 2020-01-05 1.0 NaN NaN NaN 2020-01-06 0.0 1.0 0.5 NaN 699 µs ± 43.2 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns.active)
print(drawdowns['a'].active.records)
print(drawdowns.active['a'].records)
print(drawdowns.active.records)

%timeit big_drawdowns.active
Drawdowns(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7fa004833748> of shape (6, 4)", "records_arr": "<numpy.ndarray object at 0x7fa021f50870> of shape (2,)", "col_mapper": null, "ts": "<pandas.core.frame.DataFrame object at 0x7f9ff803f898> of shape (6, 4)", "freq": "1 days" })) id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 2 0 4 5 5 5 4.0 1.0 end_val status 0 1.0 0 id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 2 0 4 5 5 5 4.0 1.0 end_val status 0 1.0 0 id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 2 0 4 5 5 5 4.0 1.0 1 5 2 2 3 4 5 3.0 1.0 end_val status 0 1.0 0 1 2.0 0 707 µs ± 42.7 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns.recovered)
print(drawdowns['a'].recovered.records)
print(drawdowns.recovered['a'].records)
print(drawdowns.recovered.records)

%timeit big_drawdowns.recovered
Drawdowns(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7fa004833748> of shape (6, 4)", "records_arr": "<numpy.ndarray object at 0x7fa021f4bad8> of shape (4,)", "col_mapper": null, "ts": "<pandas.core.frame.DataFrame object at 0x7f9ff803f898> of shape (6, 4)", "freq": "1 days" })) id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 0 0 0 1 1 2 2.0 1.0 1 1 0 2 3 3 4 3.0 1.0 end_val status 0 2.0 1 1 3.0 1 id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 0 0 0 1 1 2 2.0 1.0 1 1 0 2 3 3 4 3.0 1.0 end_val status 0 2.0 1 1 3.0 1 id col peak_idx start_idx valley_idx end_idx peak_val valley_val \ 0 0 0 0 1 1 2 2.0 1.0 1 1 0 2 3 3 4 3.0 1.0 2 3 1 1 2 2 3 2.0 1.0 3 4 1 3 4 4 5 3.0 1.0 end_val status 0 2.0 1 1 3.0 1 2 2.0 1 3 3.0 1 2.26 ms ± 266 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].active_drawdown())
print(drawdowns.active_drawdown())

%timeit big_drawdowns.active_drawdown()
-0.75 a -0.750000 b NaN c -0.333333 d NaN Name: active_drawdown, dtype: float64 2.03 ms ± 48.2 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(drawdowns['a'].active_duration())
print(drawdowns.active_duration())

%timeit big_drawdowns.active_duration()
1 days 00:00:00 a 1 days b NaT c 3 days d NaT Name: active_duration, dtype: timedelta64[ns]
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/base/array_wrapper.py:661: UserWarning: Couldn't parse the frequency of index. Pass it as `freq` or define it globally under `settings.array_wrapper`.
1.41 ms ± 65.6 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].active_recovery_return())
print(drawdowns.active_recovery_return())

%timeit big_drawdowns.active_recovery_return()
0.0 a 0.0 b NaN c 1.0 d NaN Name: active_recovery_return, dtype: float64 1.43 ms ± 43.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].active_recovery_duration())
print(drawdowns.active_recovery_duration())

%timeit big_drawdowns.active_recovery_duration()
0 days 00:00:00 a 0 days b NaT c 1 days d NaT Name: active_recovery_duration, dtype: timedelta64[ns] 1.41 ms ± 53.7 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(drawdowns['a'].stats())
%timeit big_drawdowns[0].stats(silence_warnings=True)

print(drawdowns.stats(column='a'))
%timeit big_drawdowns.stats(column=0, silence_warnings=True)

print(drawdowns.stats())
%timeit big_drawdowns.stats(silence_warnings=True)
Start 2020-01-01 00:00:00 End 2020-01-06 00:00:00 Period 6 days 00:00:00 Coverage [%] 50.0 Total Records 3 Total Recovered Drawdowns 2 Total Active Drawdowns 1 Active Drawdown [%] 75.0 Active Duration 1 days 00:00:00 Active Recovery [%] 0.0 Active Recovery Return [%] 0.0 Active Recovery Duration 0 days 00:00:00 Max Drawdown [%] 66.666667 Avg Drawdown [%] 58.333333 Max Drawdown Duration 1 days 00:00:00 Avg Drawdown Duration 1 days 00:00:00 Max Recovery Return [%] 200.0 Avg Recovery Return [%] 150.0 Max Recovery Duration 1 days 00:00:00 Avg Recovery Duration 1 days 00:00:00 Avg Recovery Duration Ratio 1.0 Name: a, dtype: object 21.6 ms ± 115 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) Start 2020-01-01 00:00:00 End 2020-01-06 00:00:00 Period 6 days 00:00:00 Coverage [%] 50.0 Total Records 3 Total Recovered Drawdowns 2 Total Active Drawdowns 1 Active Drawdown [%] 75.0 Active Duration 1 days 00:00:00 Active Recovery [%] 0.0 Active Recovery Return [%] 0.0 Active Recovery Duration 0 days 00:00:00 Max Drawdown [%] 66.666667 Avg Drawdown [%] 58.333333 Max Drawdown Duration 1 days 00:00:00 Avg Drawdown Duration 1 days 00:00:00 Max Recovery Return [%] 200.0 Avg Recovery Return [%] 150.0 Max Recovery Duration 1 days 00:00:00 Avg Recovery Duration 1 days 00:00:00 Avg Recovery Duration Ratio 1.0 Name: a, dtype: object 44.9 ms ± 2.18 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) Start 2020-01-01 00:00:00 End 2020-01-06 00:00:00 Period 6 days 00:00:00 Coverage [%] 44.444444 Total Records 1.5 Total Recovered Drawdowns 1.0 Total Active Drawdowns 0.5 Active Drawdown [%] 54.166667 Active Duration 2 days 00:00:00 Active Recovery [%] 25.0 Active Recovery Return [%] 50.0 Active Recovery Duration 0 days 12:00:00 Max Drawdown [%] 66.666667 Avg Drawdown [%] 58.333333 Max Drawdown Duration 1 days 00:00:00 Avg Drawdown Duration 1 days 00:00:00 Max Recovery Return [%] 200.0 Avg Recovery Return [%] 150.0 Max Recovery Duration 1 days 00:00:00 Avg Recovery Duration 1 days 00:00:00 Avg Recovery Duration Ratio 1.0 Name: agg_func_mean, dtype: object
/Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Object has multiple columns. Aggregating using <function mean at 0x7fa020856598>. Pass column to select a single column/group. import sys
45.2 ms ± 757 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) 
drawdowns['a'].plot().show_svg()
Image in a Jupyter notebook
drawdowns.plot(column='a', top_n=1).show_svg()
Image in a Jupyter notebook
drawdowns['a'].plots().show_svg()
Image in a Jupyter notebook

Orders

close = pd.Series([1, 2, 3, 4, 5, 6, 7, 8], index=[
    datetime(2020, 1, 1),
    datetime(2020, 1, 2),
    datetime(2020, 1, 3),
    datetime(2020, 1, 4),
    datetime(2020, 1, 5),
    datetime(2020, 1, 6),
    datetime(2020, 1, 7),
    datetime(2020, 1, 8)
]).vbt.tile(4, keys=['a', 'b', 'c', 'd'])
print(close)

big_close = pd.DataFrame(np.random.uniform(1, 10, size=(1000, 1000)))
a b c d 2020-01-01 1 1 1 1 2020-01-02 2 2 2 2 2020-01-03 3 3 3 3 2020-01-04 4 4 4 4 2020-01-05 5 5 5 5 2020-01-06 6 6 6 6 2020-01-07 7 7 7 7 2020-01-08 8 8 8 8 
from vectorbt.portfolio.enums import order_dt

records_arr = np.asarray([
    (0, 0, 0, 1. , 1., 0.01 , 0), (1, 0, 1, 0.1, 2., 0.002, 0),
    (2, 0, 2, 1. , 3., 0.03 , 1), (3, 0, 3, 0.1, 4., 0.004, 1),
    (4, 0, 5, 1. , 6., 0.06 , 0), (5, 0, 6, 1. , 7., 0.07 , 1),
    (6, 0, 7, 2. , 8., 0.16 , 0), (7, 1, 0, 1. , 1., 0.01 , 1),
    (8, 1, 1, 0.1, 2., 0.002, 1), (9, 1, 2, 1. , 3., 0.03 , 0),
    (10, 1, 3, 0.1, 4., 0.004, 0), (11, 1, 5, 1. , 6., 0.06 , 1),
    (12, 1, 6, 1. , 7., 0.07 , 0), (13, 1, 7, 2. , 8., 0.16 , 1),
    (14, 2, 0, 1. , 1., 0.01 , 0), (15, 2, 1, 0.1, 2., 0.002, 0),
    (16, 2, 2, 1. , 3., 0.03 , 1), (17, 2, 3, 0.1, 4., 0.004, 1),
    (18, 2, 5, 1. , 6., 0.06 , 0), (19, 2, 6, 2. , 7., 0.14 , 1),
    (20, 2, 7, 2. , 8., 0.16 , 0)
], dtype=order_dt)
print(records_arr.shape)

wrapper = vbt.ArrayWrapper.from_obj(close, freq='1 days')
orders = vbt.Orders(wrapper, records_arr, close)
orders_grouped = vbt.Orders(wrapper.regroup(group_by), records_arr, close)
(21,) 
big_records_arr = np.asarray(list(zip(*(
    np.arange(1000000),
    np.tile(np.arange(1000), 1000),
    np.repeat(np.arange(1000), 1000),
    np.full(1000000, 10),
    np.random.uniform(1, 10, size=1000000),
    np.full(1000000, 1),
    np.full(1000000, 1)
))), dtype=order_dt)
big_records_arr['side'][::2] = 0
print(big_records_arr.shape)

big_wrapper = vbt.ArrayWrapper.from_obj(big_close, freq='1 days')
big_orders = vbt.Orders(big_wrapper, big_records_arr, big_close)
big_orders_grouped = vbt.Orders(big_wrapper.copy(group_by=big_group_by), big_records_arr, big_close)
(1000000,) 
print(orders.records)
print(orders.close)

print(orders['a'].records)
print(orders['a'].close)

%timeit big_orders.iloc[0]
%timeit big_orders.iloc[:]

%timeit big_orders_grouped.iloc[0]
%timeit big_orders_grouped.iloc[:]
id col idx size price fees side 0 0 0 0 1.0 1.0 0.010 0 1 1 0 1 0.1 2.0 0.002 0 2 2 0 2 1.0 3.0 0.030 1 3 3 0 3 0.1 4.0 0.004 1 4 4 0 5 1.0 6.0 0.060 0 5 5 0 6 1.0 7.0 0.070 1 6 6 0 7 2.0 8.0 0.160 0 7 7 1 0 1.0 1.0 0.010 1 8 8 1 1 0.1 2.0 0.002 1 9 9 1 2 1.0 3.0 0.030 0 10 10 1 3 0.1 4.0 0.004 0 11 11 1 5 1.0 6.0 0.060 1 12 12 1 6 1.0 7.0 0.070 0 13 13 1 7 2.0 8.0 0.160 1 14 14 2 0 1.0 1.0 0.010 0 15 15 2 1 0.1 2.0 0.002 0 16 16 2 2 1.0 3.0 0.030 1 17 17 2 3 0.1 4.0 0.004 1 18 18 2 5 1.0 6.0 0.060 0 19 19 2 6 2.0 7.0 0.140 1 20 20 2 7 2.0 8.0 0.160 0 a b c d 2020-01-01 1 1 1 1 2020-01-02 2 2 2 2 2020-01-03 3 3 3 3 2020-01-04 4 4 4 4 2020-01-05 5 5 5 5 2020-01-06 6 6 6 6 2020-01-07 7 7 7 7 2020-01-08 8 8 8 8 id col idx size price fees side 0 0 0 0 1.0 1.0 0.010 0 1 1 0 1 0.1 2.0 0.002 0 2 2 0 2 1.0 3.0 0.030 1 3 3 0 3 0.1 4.0 0.004 1 4 4 0 5 1.0 6.0 0.060 0 5 5 0 6 1.0 7.0 0.070 1 6 6 0 7 2.0 8.0 0.160 0 2020-01-01 1 2020-01-02 2 2020-01-03 3 2020-01-04 4 2020-01-05 5 2020-01-06 6 2020-01-07 7 2020-01-08 8 Name: a, dtype: int64 1.06 ms ± 17.4 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 12.5 ms ± 2.47 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) 13.5 ms ± 1.52 ms per loop (mean ± std. dev. of 7 runs, 100 loops each) 13 ms ± 4.86 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) 
print(orders.records_readable)
Order Id Column Timestamp Size Price Fees Side 0 0 a 2020-01-01 1.0 1.0 0.010 Buy 1 1 a 2020-01-02 0.1 2.0 0.002 Buy 2 2 a 2020-01-03 1.0 3.0 0.030 Sell 3 3 a 2020-01-04 0.1 4.0 0.004 Sell 4 4 a 2020-01-06 1.0 6.0 0.060 Buy 5 5 a 2020-01-07 1.0 7.0 0.070 Sell 6 6 a 2020-01-08 2.0 8.0 0.160 Buy 7 7 b 2020-01-01 1.0 1.0 0.010 Sell 8 8 b 2020-01-02 0.1 2.0 0.002 Sell 9 9 b 2020-01-03 1.0 3.0 0.030 Buy 10 10 b 2020-01-04 0.1 4.0 0.004 Buy 11 11 b 2020-01-06 1.0 6.0 0.060 Sell 12 12 b 2020-01-07 1.0 7.0 0.070 Buy 13 13 b 2020-01-08 2.0 8.0 0.160 Sell 14 14 c 2020-01-01 1.0 1.0 0.010 Buy 15 15 c 2020-01-02 0.1 2.0 0.002 Buy 16 16 c 2020-01-03 1.0 3.0 0.030 Sell 17 17 c 2020-01-04 0.1 4.0 0.004 Sell 18 18 c 2020-01-06 1.0 6.0 0.060 Buy 19 19 c 2020-01-07 2.0 7.0 0.140 Sell 20 20 c 2020-01-08 2.0 8.0 0.160 Buy 
print(orders.buy)
print(orders['a'].buy.records)
print(orders.buy['a'].records)
print(orders.buy.records)

%timeit big_orders.buy
Orders(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7f9f780709d0> of shape (11,)", "col_mapper": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col idx size price fees side 0 0 0 0 1.0 1.0 0.010 0 1 1 0 1 0.1 2.0 0.002 0 2 4 0 5 1.0 6.0 0.060 0 3 6 0 7 2.0 8.0 0.160 0 id col idx size price fees side 0 0 0 0 1.0 1.0 0.010 0 1 1 0 1 0.1 2.0 0.002 0 2 4 0 5 1.0 6.0 0.060 0 3 6 0 7 2.0 8.0 0.160 0 id col idx size price fees side 0 0 0 0 1.0 1.0 0.010 0 1 1 0 1 0.1 2.0 0.002 0 2 4 0 5 1.0 6.0 0.060 0 3 6 0 7 2.0 8.0 0.160 0 4 9 1 2 1.0 3.0 0.030 0 5 10 1 3 0.1 4.0 0.004 0 6 12 1 6 1.0 7.0 0.070 0 7 14 2 0 1.0 1.0 0.010 0 8 15 2 1 0.1 2.0 0.002 0 9 18 2 5 1.0 6.0 0.060 0 10 20 2 7 2.0 8.0 0.160 0 11.1 ms ± 387 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(orders.sell)
print(orders['a'].sell.records)
print(orders.sell['a'].records)
print(orders.sell.records)

%timeit big_orders.sell
Orders(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7fa0223c06b8> of shape (10,)", "col_mapper": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col idx size price fees side 0 2 0 2 1.0 3.0 0.030 1 1 3 0 3 0.1 4.0 0.004 1 2 5 0 6 1.0 7.0 0.070 1 id col idx size price fees side 0 2 0 2 1.0 3.0 0.030 1 1 3 0 3 0.1 4.0 0.004 1 2 5 0 6 1.0 7.0 0.070 1 id col idx size price fees side 0 2 0 2 1.0 3.0 0.030 1 1 3 0 3 0.1 4.0 0.004 1 2 5 0 6 1.0 7.0 0.070 1 3 7 1 0 1.0 1.0 0.010 1 4 8 1 1 0.1 2.0 0.002 1 5 11 1 5 1.0 6.0 0.060 1 6 13 1 7 2.0 8.0 0.160 1 7 16 2 2 1.0 3.0 0.030 1 8 17 2 3 0.1 4.0 0.004 1 9 19 2 6 2.0 7.0 0.140 1 10.7 ms ± 117 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(orders['a'].stats())
%timeit big_orders[0].stats(silence_warnings=True)

print(orders.stats(column='a'))
%timeit big_orders.stats(column=0, silence_warnings=True)

print(orders.stats())
%timeit big_orders.stats(silence_warnings=True)
Start 2020-01-01 00:00:00 End 2020-01-08 00:00:00 Period 8 days 00:00:00 Total Records 7 Total Buy Orders 4 Total Sell Orders 3 Min Size 0.1 Max Size 2.0 Avg Size 0.885714 Avg Buy Size 1.025 Avg Sell Size 0.7 Avg Buy Price 4.25 Avg Sell Price 4.666667 Total Fees 0.336 Min Fees 0.002 Max Fees 0.16 Avg Fees 0.048 Avg Buy Fees 0.058 Avg Sell Fees 0.034667 Name: a, dtype: object 14 ms ± 171 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) Start 2020-01-01 00:00:00 End 2020-01-08 00:00:00 Period 8 days 00:00:00 Total Records 7 Total Buy Orders 4 Total Sell Orders 3 Min Size 0.1 Max Size 2.0 Avg Size 0.885714 Avg Buy Size 1.025 Avg Sell Size 0.7 Avg Buy Price 4.25 Avg Sell Price 4.666667 Total Fees 0.336 Min Fees 0.002 Max Fees 0.16 Avg Fees 0.048 Avg Buy Fees 0.058 Avg Sell Fees 0.034667 Name: a, dtype: object 85.6 ms ± 3.18 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) Start 2020-01-01 00:00:00 End 2020-01-08 00:00:00 Period 8 days 00:00:00 Total Records 5.25 Total Buy Orders 2.75 Total Sell Orders 2.5 Min Size 0.1 Max Size 2.0 Avg Size 0.933333 Avg Buy Size 0.916667 Avg Sell Size 0.919444 Avg Buy Price 4.388889 Avg Sell Price 4.527778 Total Fees 0.2695 Min Fees 0.002 Max Fees 0.16 Avg Fees 0.051333 Avg Buy Fees 0.050222 Avg Sell Fees 0.050222 Name: agg_func_mean, dtype: object
/Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Object has multiple columns. Aggregating using <function mean at 0x7fa020856598>. Pass column to select a single column/group. import sys
84.7 ms ± 2.13 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) 
orders['a'].plot().show_svg()
Image in a Jupyter notebook
orders.plot(column='a').show_svg()
Image in a Jupyter notebook
orders.plots(column='a').show_svg()
Image in a Jupyter notebook

Trades

trades = vbt.ExitTrades.from_orders(orders)
trades_grouped = vbt.ExitTrades.from_orders(orders_grouped)
print(trades.values.shape)
(13,) 
big_trades = vbt.ExitTrades.from_orders(big_orders)
big_trades_grouped = vbt.ExitTrades.from_orders(big_orders_grouped)
print(big_trades.values.shape)
(1000,) 
%timeit vbt.ExitTrades.from_orders(big_orders)
4.55 ms ± 46.1 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(trades.records)
print(trades.close)

print(trades['a'].records)
print(trades['a'].close)

%timeit big_trades.iloc[0]
%timeit big_trades.iloc[:]

%timeit big_trades_grouped.iloc[0]
%timeit big_trades_grouped.iloc[:]
id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 3 0 2.0 7 8.000000 0.160000 7 8.0 4 4 1 1.0 0 1.090909 0.010909 2 3.0 5 5 1 0.1 0 1.090909 0.001091 3 4.0 6 6 1 1.0 5 6.000000 0.060000 6 7.0 7 7 1 2.0 7 8.000000 0.160000 7 8.0 8 8 2 1.0 0 1.090909 0.010909 2 3.0 9 9 2 0.1 0 1.090909 0.001091 3 4.0 10 10 2 1.0 5 6.000000 0.060000 6 7.0 11 11 2 1.0 6 7.000000 0.070000 7 8.0 12 12 2 1.0 7 8.000000 0.080000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.712500 0 1 0 1 0.004 0.285818 2.620000 0 1 0 2 0.070 0.870000 0.145000 0 1 1 3 0.000 -0.160000 -0.010000 0 0 2 4 0.030 -1.950000 -1.787500 1 1 3 5 0.004 -0.296000 -2.713333 1 1 3 6 0.070 -1.130000 -0.188333 1 1 4 7 0.000 -0.160000 -0.010000 1 0 5 8 0.030 1.868182 1.712500 0 1 6 9 0.004 0.285818 2.620000 0 1 6 10 0.070 0.870000 0.145000 0 1 7 11 0.080 -1.150000 -0.164286 1 1 8 12 0.000 -0.080000 -0.010000 0 0 9 a b c d 2020-01-01 1 1 1 1 2020-01-02 2 2 2 2 2020-01-03 3 3 3 3 2020-01-04 4 4 4 4 2020-01-05 5 5 5 5 2020-01-06 6 6 6 6 2020-01-07 7 7 7 7 2020-01-08 8 8 8 8 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 3 0 2.0 7 8.000000 0.160000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 3 0.000 -0.160000 -0.0100 0 0 2 2020-01-01 1 2020-01-02 2 2020-01-03 3 2020-01-04 4 2020-01-05 5 2020-01-06 6 2020-01-07 7 2020-01-08 8 Name: a, dtype: int64 1.56 ms ± 288 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 3.37 ms ± 567 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 2.84 ms ± 218 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 3.65 ms ± 225 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(trades.records_readable)
Trade Id Column Size Entry Timestamp Avg Entry Price Entry Fees \ 0 0 a 1.0 2020-01-01 1.090909 0.010909 1 1 a 0.1 2020-01-01 1.090909 0.001091 2 2 a 1.0 2020-01-06 6.000000 0.060000 3 3 a 2.0 2020-01-08 8.000000 0.160000 4 4 b 1.0 2020-01-01 1.090909 0.010909 5 5 b 0.1 2020-01-01 1.090909 0.001091 6 6 b 1.0 2020-01-06 6.000000 0.060000 7 7 b 2.0 2020-01-08 8.000000 0.160000 8 8 c 1.0 2020-01-01 1.090909 0.010909 9 9 c 0.1 2020-01-01 1.090909 0.001091 10 10 c 1.0 2020-01-06 6.000000 0.060000 11 11 c 1.0 2020-01-07 7.000000 0.070000 12 12 c 1.0 2020-01-08 8.000000 0.080000 Exit Timestamp Avg Exit Price Exit Fees PnL Return Direction \ 0 2020-01-03 3.0 0.030 1.868182 1.712500 Long 1 2020-01-04 4.0 0.004 0.285818 2.620000 Long 2 2020-01-07 7.0 0.070 0.870000 0.145000 Long 3 2020-01-08 8.0 0.000 -0.160000 -0.010000 Long 4 2020-01-03 3.0 0.030 -1.950000 -1.787500 Short 5 2020-01-04 4.0 0.004 -0.296000 -2.713333 Short 6 2020-01-07 7.0 0.070 -1.130000 -0.188333 Short 7 2020-01-08 8.0 0.000 -0.160000 -0.010000 Short 8 2020-01-03 3.0 0.030 1.868182 1.712500 Long 9 2020-01-04 4.0 0.004 0.285818 2.620000 Long 10 2020-01-07 7.0 0.070 0.870000 0.145000 Long 11 2020-01-08 8.0 0.080 -1.150000 -0.164286 Short 12 2020-01-08 8.0 0.000 -0.080000 -0.010000 Long Status Position Id 0 Closed 0 1 Closed 0 2 Closed 1 3 Open 2 4 Closed 3 5 Closed 3 6 Closed 4 7 Open 5 8 Closed 6 9 Closed 6 10 Closed 7 11 Closed 8 12 Open 9 
print(trades['a'].count())
print(trades.count())
%timeit big_trades.count()

print(trades_grouped.count())
%timeit big_trades_grouped.count()
4 a 4 b 4 c 5 d 0 Name: count, dtype: int64 49.4 µs ± 504 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each) group first 8 second 5 Name: count, dtype: int64 49 µs ± 303 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each) 
print(trades.winning)
print(trades['a'].winning.records)
print(trades.winning['a'].records)
print(trades.winning.records)

%timeit big_trades.winning
Trades(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7f9f7cf155b0> of shape (6,)", "col_mapper": null, "ts": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 8 2 1.0 0 1.090909 0.010909 2 3.0 4 9 2 0.1 0 1.090909 0.001091 3 4.0 5 10 2 1.0 5 6.000000 0.060000 6 7.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 3 0.030 1.868182 1.7125 0 1 6 4 0.004 0.285818 2.6200 0 1 6 5 0.070 0.870000 0.1450 0 1 7 756 µs ± 11 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades.losing)
print(trades['a'].losing.records)
print(trades.losing['a'].records)
print(trades.losing.records)

%timeit big_trades.losing
Trades(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7fa007ca3a28> of shape (7,)", "col_mapper": null, "ts": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 3 0 2.0 7 8.0 0.16 7 8.0 exit_fees pnl return direction status position_id 0 0.0 -0.16 -0.01 0 0 2 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 3 0 2.0 7 8.0 0.16 7 8.0 exit_fees pnl return direction status position_id 0 0.0 -0.16 -0.01 0 0 2 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 3 0 2.0 7 8.000000 0.160000 7 8.0 1 4 1 1.0 0 1.090909 0.010909 2 3.0 2 5 1 0.1 0 1.090909 0.001091 3 4.0 3 6 1 1.0 5 6.000000 0.060000 6 7.0 4 7 1 2.0 7 8.000000 0.160000 7 8.0 5 11 2 1.0 6 7.000000 0.070000 7 8.0 6 12 2 1.0 7 8.000000 0.080000 7 8.0 exit_fees pnl return direction status position_id 0 0.000 -0.160 -0.010000 0 0 2 1 0.030 -1.950 -1.787500 1 1 3 2 0.004 -0.296 -2.713333 1 1 3 3 0.070 -1.130 -0.188333 1 1 4 4 0.000 -0.160 -0.010000 1 0 5 5 0.080 -1.150 -0.164286 1 1 8 6 0.000 -0.080 -0.010000 0 0 9 751 µs ± 8.53 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades['a'].winning_streak.to_pd(ignore_index=True))
print(trades.winning_streak.to_pd(ignore_index=True))

%timeit big_trades.winning_streak
0 1 1 2 2 3 3 0 Name: a, dtype: int64 a b c d 0 1.0 0.0 1.0 NaN 1 2.0 0.0 2.0 NaN 2 3.0 0.0 3.0 NaN 3 0.0 0.0 0.0 NaN 4 NaN NaN 0.0 NaN 565 µs ± 7.54 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades['a'].losing_streak.to_pd(ignore_index=True))
print(trades.losing_streak.to_pd(ignore_index=True))

%timeit big_trades.losing_streak
0 0 1 0 2 0 3 1 Name: a, dtype: int64 a b c d 0 0.0 1.0 0.0 NaN 1 0.0 2.0 0.0 NaN 2 0.0 3.0 0.0 NaN 3 1.0 4.0 1.0 NaN 4 NaN NaN 2.0 NaN 551 µs ± 5.51 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades['a'].win_rate())
print(trades.win_rate())
%timeit big_trades.win_rate()

print(trades.win_rate(group_by=group_by))
%timeit big_trades.win_rate(group_by=big_group_by)
0.75 a 0.75 b 0.00 c 0.60 d NaN Name: win_rate, dtype: float64
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:530: RuntimeWarning: invalid value encountered in true_divide
951 µs ± 3.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first 0.375 second 0.600 Name: win_rate, dtype: float64 5.47 ms ± 21.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(trades['a'].profit_factor())
print(trades.profit_factor())
%timeit big_trades.profit_factor()

print(trades_grouped.profit_factor())
%timeit big_trades_grouped.profit_factor()
18.9 a 18.900000 b 0.000000 c 2.458537 d NaN Name: profit_factor, dtype: float64
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: divide by zero encountered in true_divide
2.65 ms ± 136 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) group first 0.818182 second 2.458537 Name: profit_factor, dtype: float64 2.57 ms ± 65.5 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(trades['a'].expectancy())
print(trades.expectancy())
%timeit big_trades.expectancy()

print(trades_grouped.expectancy())
%timeit big_trades_grouped.expectancy()
0.716 a 0.7160 b -0.8840 c 0.3588 d NaN Name: expectancy, dtype: float64
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:530: RuntimeWarning: invalid value encountered in true_divide
3.55 ms ± 34.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) group first -0.0840 second 0.3588 Name: expectancy, dtype: float64 3.54 ms ± 37.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 
print(trades['a'].sqn())
print(trades.sqn())
%timeit big_trades.sqn()

print(trades_grouped.sqn())
%timeit big_trades_grouped.sqn()
1.634155521947584 a 1.634156 b -2.130073 c 0.716604 d NaN Name: sqn, dtype: float64 1.32 ms ± 11.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) group first -0.204047 second 0.716604 Name: sqn, dtype: float64 924 µs ± 9.78 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades.long)
print(trades['a'].long.records)
print(trades.long['a'].records)
print(trades.long.records)

%timeit big_trades.long
Trades(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7f9f68734be0> of shape (8,)", "col_mapper": null, "ts": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 3 0 2.0 7 8.000000 0.160000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 3 0.000 -0.160000 -0.0100 0 0 2 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 3 0 2.0 7 8.000000 0.160000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 3 0.000 -0.160000 -0.0100 0 0 2 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 3 0 2.0 7 8.000000 0.160000 7 8.0 4 8 2 1.0 0 1.090909 0.010909 2 3.0 5 9 2 0.1 0 1.090909 0.001091 3 4.0 6 10 2 1.0 5 6.000000 0.060000 6 7.0 7 12 2 1.0 7 8.000000 0.080000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 3 0.000 -0.160000 -0.0100 0 0 2 4 0.030 1.868182 1.7125 0 1 6 5 0.004 0.285818 2.6200 0 1 6 6 0.070 0.870000 0.1450 0 1 7 7 0.000 -0.080000 -0.0100 0 0 9 764 µs ± 12.7 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades.short)
print(trades['a'].short.records)
print(trades.short['a'].records)
print(trades.short.records)

%timeit big_trades.short
Trades(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7f9f7a2d8608> of shape (5,)", "col_mapper": null, "ts": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) Empty DataFrame Columns: [id, col, size, entry_idx, entry_price, entry_fees, exit_idx, exit_price, exit_fees, pnl, return, direction, status, position_id] Index: [] Empty DataFrame Columns: [id, col, size, entry_idx, entry_price, entry_fees, exit_idx, exit_price, exit_fees, pnl, return, direction, status, position_id] Index: [] id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 4 1 1.0 0 1.090909 0.010909 2 3.0 1 5 1 0.1 0 1.090909 0.001091 3 4.0 2 6 1 1.0 5 6.000000 0.060000 6 7.0 3 7 1 2.0 7 8.000000 0.160000 7 8.0 4 11 2 1.0 6 7.000000 0.070000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 -1.950 -1.787500 1 1 3 1 0.004 -0.296 -2.713333 1 1 3 2 0.070 -1.130 -0.188333 1 1 4 3 0.000 -0.160 -0.010000 1 0 5 4 0.080 -1.150 -0.164286 1 1 8 762 µs ± 9.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades.open)
print(trades['a'].open.records)
print(trades.open['a'].records)
print(trades.open.records)

%timeit big_trades.open
Trades(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7f9f7cd48978> of shape (3,)", "col_mapper": null, "ts": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 3 0 2.0 7 8.0 0.16 7 8.0 exit_fees pnl return direction status position_id 0 0.0 -0.16 -0.01 0 0 2 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 3 0 2.0 7 8.0 0.16 7 8.0 exit_fees pnl return direction status position_id 0 0.0 -0.16 -0.01 0 0 2 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 3 0 2.0 7 8.0 0.16 7 8.0 1 7 1 2.0 7 8.0 0.16 7 8.0 2 12 2 1.0 7 8.0 0.08 7 8.0 exit_fees pnl return direction status position_id 0 0.0 -0.16 -0.01 0 0 2 1 0.0 -0.16 -0.01 1 0 5 2 0.0 -0.08 -0.01 0 0 9 777 µs ± 10.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades.closed)
print(trades['a'].closed.records)
print(trades.closed['a'].records)
print(trades.closed.records)

%timeit big_trades.closed
Trades(**Config({ "wrapper": "<vectorbt.base.array_wrapper.ArrayWrapper object at 0x7f9f78606d30> of shape (8, 4)", "records_arr": "<numpy.ndarray object at 0x7f9f7a4e3608> of shape (10,)", "col_mapper": null, "ts": null, "close": "<pandas.core.frame.DataFrame object at 0x7f9f782a4390> of shape (8, 4)" })) id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.7125 0 1 0 1 0.004 0.285818 2.6200 0 1 0 2 0.070 0.870000 0.1450 0 1 1 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.0 0 1.090909 0.010909 2 3.0 1 1 0 0.1 0 1.090909 0.001091 3 4.0 2 2 0 1.0 5 6.000000 0.060000 6 7.0 3 4 1 1.0 0 1.090909 0.010909 2 3.0 4 5 1 0.1 0 1.090909 0.001091 3 4.0 5 6 1 1.0 5 6.000000 0.060000 6 7.0 6 8 2 1.0 0 1.090909 0.010909 2 3.0 7 9 2 0.1 0 1.090909 0.001091 3 4.0 8 10 2 1.0 5 6.000000 0.060000 6 7.0 9 11 2 1.0 6 7.000000 0.070000 7 8.0 exit_fees pnl return direction status position_id 0 0.030 1.868182 1.712500 0 1 0 1 0.004 0.285818 2.620000 0 1 0 2 0.070 0.870000 0.145000 0 1 1 3 0.030 -1.950000 -1.787500 1 1 3 4 0.004 -0.296000 -2.713333 1 1 3 5 0.070 -1.130000 -0.188333 1 1 4 6 0.030 1.868182 1.712500 0 1 6 7 0.004 0.285818 2.620000 0 1 6 8 0.070 0.870000 0.145000 0 1 7 9 0.080 -1.150000 -0.164286 1 1 8 759 µs ± 11.4 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(trades['a'].stats())
%timeit big_trades[0].stats(silence_warnings=True)

print(trades.stats(column='a'))
%timeit big_trades.stats(column=0, silence_warnings=True)

print(trades.stats())
%timeit big_trades.stats(silence_warnings=True)
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: divide by zero encountered in true_divide
Start 2020-01-01 00:00:00 End 2020-01-08 00:00:00 Period 8 days 00:00:00 First Trade Start 2020-01-01 00:00:00 Last Trade End 2020-01-08 00:00:00 Coverage 5 days 00:00:00 Overlap Coverage 2 days 00:00:00 Total Records 4 Total Long Trades 4 Total Short Trades 0 Total Closed Trades 3 Total Open Trades 1 Open Trade PnL -0.16 Win Rate [%] 100.0 Max Win Streak 3 Max Loss Streak 0 Best Trade [%] 262.0 Worst Trade [%] 14.5 Avg Winning Trade [%] 149.25 Avg Losing Trade [%] NaN Avg Winning Trade Duration 2 days 00:00:00 Avg Losing Trade Duration NaT Profit Factor inf Expectancy 1.008 SQN 2.181955 Name: a, dtype: object
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:530: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:530: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: invalid value encountered in true_divide
42.1 ms ± 2.02 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) Start 2020-01-01 00:00:00 End 2020-01-08 00:00:00 Period 8 days 00:00:00 First Trade Start 2020-01-01 00:00:00 Last Trade End 2020-01-08 00:00:00 Coverage 5 days 00:00:00 Overlap Coverage 2 days 00:00:00 Total Records 4 Total Long Trades 4 Total Short Trades 0 Total Closed Trades 3 Total Open Trades 1 Open Trade PnL -0.16 Win Rate [%] 100.0 Max Win Streak 3 Max Loss Streak 0 Best Trade [%] 262.0 Worst Trade [%] 14.5 Avg Winning Trade [%] 149.25 Avg Losing Trade [%] NaN Avg Winning Trade Duration 2 days 00:00:00 Avg Losing Trade Duration NaT Profit Factor inf Expectancy 1.008 SQN 2.181955 Name: a, dtype: object
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:530: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: divide by zero encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: invalid value encountered in true_divide
44.4 ms ± 1.17 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) Start 2020-01-01 00:00:00 End 2020-01-08 00:00:00 Period 8 days 00:00:00 First Trade Start 2020-01-01 00:00:00 Last Trade End 2020-01-08 00:00:00 Coverage 5 days 08:00:00 Overlap Coverage 2 days 00:00:00 Total Records 3.25 Total Long Trades 2.0 Total Short Trades 1.25 Total Closed Trades 2.5 Total Open Trades 0.75 Open Trade PnL -0.1 Win Rate [%] 58.333333 Max Win Streak 2.0 Max Loss Streak 1.333333 Best Trade [%] 168.388889 Worst Trade [%] -91.087302 Avg Winning Trade [%] 149.25 Avg Losing Trade [%] -86.367063 Avg Winning Trade Duration 2 days 00:00:00 Avg Losing Trade Duration 1 days 12:00:00 Profit Factor inf Expectancy 0.117056 SQN 0.189316 Name: agg_func_mean, dtype: object
/Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:530: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: divide by zero encountered in true_divide /Users/olegpolakow/Documents/SourceTree/vectorbt/vectorbt/portfolio/trades.py:544: RuntimeWarning: invalid value encountered in true_divide /Users/olegpolakow/miniconda3/lib/python3.7/site-packages/ipykernel_launcher.py:7: UserWarning: Object has multiple columns. Aggregating using <function mean at 0x7fa020856598>. Pass column to select a single column/group.
44.9 ms ± 1.26 ms per loop (mean ± std. dev. of 7 runs, 10 loops each) 
trades['c'].plot().show_svg()
Image in a Jupyter notebook
trades['c'].plot_pnl().show_svg()
Image in a Jupyter notebook
trades.plots(column='c').show_svg()
Image in a Jupyter notebook

Positions

positions = vbt.Positions.from_trades(trades)
positions_grouped = vbt.Positions.from_trades(trades_grouped)
print(positions.values.shape)
(10,) 
big_positions = vbt.Positions.from_trades(big_trades)
big_positions_grouped = vbt.Positions.from_trades(big_trades_grouped)
print(big_positions.values.shape)
(1000,) 
%timeit vbt.Positions.from_trades(big_trades)
724 µs ± 12.6 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 
print(positions.records)
print(positions.close)

print(positions['a'].records)
print(positions['a'].close)

%timeit big_positions.iloc[0]
%timeit big_positions.iloc[:]

%timeit big_positions_grouped.iloc[0]
%timeit big_positions_grouped.iloc[:]
id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.1 0 1.090909 0.012 3 3.090909 1 1 0 1.0 5 6.000000 0.060 6 7.000000 2 2 0 2.0 7 8.000000 0.160 7 8.000000 3 3 1 1.1 0 1.090909 0.012 3 3.090909 4 4 1 1.0 5 6.000000 0.060 6 7.000000 5 5 1 2.0 7 8.000000 0.160 7 8.000000 6 6 2 1.1 0 1.090909 0.012 3 3.090909 7 7 2 1.0 5 6.000000 0.060 6 7.000000 8 8 2 1.0 6 7.000000 0.070 7 8.000000 9 9 2 1.0 7 8.000000 0.080 7 8.000000 exit_fees pnl return direction status 0 0.034 2.154 1.795000 0 1 1 0.070 0.870 0.145000 0 1 2 0.000 -0.160 -0.010000 0 0 3 0.034 -2.246 -1.871667 1 1 4 0.070 -1.130 -0.188333 1 1 5 0.000 -0.160 -0.010000 1 0 6 0.034 2.154 1.795000 0 1 7 0.070 0.870 0.145000 0 1 8 0.080 -1.150 -0.164286 1 1 9 0.000 -0.080 -0.010000 0 0 a b c d 2020-01-01 1 1 1 1 2020-01-02 2 2 2 2 2020-01-03 3 3 3 3 2020-01-04 4 4 4 4 2020-01-05 5 5 5 5 2020-01-06 6 6 6 6 2020-01-07 7 7 7 7 2020-01-08 8 8 8 8 id col size entry_idx entry_price entry_fees exit_idx exit_price \ 0 0 0 1.1 0 1.090909 0.012 3 3.090909 1 1 0 1.0 5 6.000000 0.060 6 7.000000 2 2 0 2.0 7 8.000000 0.160 7 8.000000 exit_fees pnl return direction status 0 0.034 2.154 1.795 0 1 1 0.070 0.870 0.145 0 1 2 0.000 -0.160 -0.010 0 0 2020-01-01 1 2020-01-02 2 2020-01-03 3 2020-01-04 4 2020-01-05 5 2020-01-06 6 2020-01-07 7 2020-01-08 8 Name: a, dtype: int64 1.4 ms ± 43.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each) 3.06 ms ± 757 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 2.52 ms ± 53.5 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 3.22 ms ± 89.3 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)