Detail: IndexHierarchy: Method

IndexHierarchy.__array__(dtype=None)

Support the __array__ interface, returning an array of values.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.__array__()
[[    0  1024    32]
 [    1 -2048    32]
 [    1  1024    32]]

IndexHierarchy.__array_ufunc__(ufunc, method, *args, **kwargs)

Support for NumPy elements or arrays on the left hand of binary operators.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> np.array((0, 1, 0)) * ih
[[    0  1024     0]
 [    0 -2048     0]
 [    0  1024     0]]

IndexHierarchy.__bool__()

Raises ValueError to prohibit ambiguous use of truthy evaluation.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> bool(ih)
ErrorNotTruthy('The truth value of a container is ambiguous. For a truthy indicator of non-empty status, use the `size` attribute.')

IndexHierarchy.__copy__()[source]

Return a shallow copy of this IndexHierarchy.

>>> import copy
>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> copy.copy(ih)
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>

IndexHierarchy.__deepcopy__(memo)[source]

Return a deep copy of this IndexHierarchy.

>>> import copy
>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> copy.deepcopy(ih)
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>

IndexHierarchy.__len__()[source]

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> len(ih)
3

IndexHierarchy.all(axis=0, skipna=True, out=None)

Logical and over values along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((False, True, True), (True, True, True), (False, True, False)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
False                             True   True
True                              True   True
False                             True   False
<bool>                            <bool> <bool>
>>> ih.all()
NotImplementedError('<function ufunc_all at 0x7f65bb2e4860> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.any(axis=0, skipna=True, out=None)

Logical or over values along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((False, True, True), (True, True, True), (False, True, False)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
False                             True   True
True                              True   True
False                             True   False
<bool>                            <bool> <bool>
>>> ih.any()
NotImplementedError('<function ufunc_any at 0x7f65bb2e4900> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.astype[key](dtypes, *, consolidate_blocks)

astype

Retype one or more depths. Can be used as as function to retype the entire IndexHierarchy; alternatively, a __getitem__ interface permits retyping selected depths.

Parameters:: dtype – A value suitable for specyfying a NumPy dtype, such as a Python type (float), NumPy array protocol strings (‘f8’), or a dtype instance.

InterfaceIndexHierarchyAsType.__getitem__(key)[source]

Selector of columns by label.

Parameters:: key – A loc selector, either a label, a list of labels, a slice of labels, or a Boolean array.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, '1517-04-01'), ('a', 2048, '1789-12-31'), ('b', 0, '1620-11-21')), index_constructors=(sf.Index, sf.Index, sf.IndexDate), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
a                                 1024    1517-04-01
a                                 2048    1789-12-31
b                                 0       1620-11-21
<<U1>                             <int64> <datetime64[D]>
>>> ih.astype[1](bool)
<IndexHierarchy: ('x', 'y', 'z')>
a                                 True   1517-04-01
a                                 True   1789-12-31
b                                 False  1620-11-21
<<U1>                             <bool> <datetime64[D]>

IndexHierarchy.astype(dtype, *, consolidate_blocks)

astype

Retype one or more depths. Can be used as as function to retype the entire IndexHierarchy; alternatively, a __getitem__ interface permits retyping selected depths.

Parameters:: dtype – A value suitable for specyfying a NumPy dtype, such as a Python type (float), NumPy array protocol strings (‘f8’), or a dtype instance.

InterfaceIndexHierarchyAsType.__call__(dtype, *, consolidate_blocks=False)[source]: Apply a single dtype to all columns.

>>> ih = sf.IndexHierarchy.from_labels(((False, True, True), (True, True, True), (False, True, False)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
False                             True   True
True                              True   True
False                             True   False
<bool>                            <bool> <bool>
>>> ih.astype(str)
<IndexHierarchy: ('x', 'y', 'z')>
False                             True  True
True                              True  True
False                             True  False
<<U5>                             <<U5> <<U5>

IndexHierarchy.copy()[source]

Return a shallow copy of this IndexHierarchy.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.copy()
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>

IndexHierarchy.cumprod(axis=0, skipna=True)

Return the cumulative product over the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.cumprod()
NotImplementedError('<function cumprod at 0x7f65f318a5c0> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.cumsum(axis=0, skipna=True)

Return the cumulative sum over the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.cumsum()
NotImplementedError('<function cumsum at 0x7f65f3189e40> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.difference(*others)[source]

Perform difference with another Index, container, or NumPy array. Retains order.

>>> ih1 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih1
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih2 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('b', 1024, True)), name='y')
>>> ih2
<IndexHierarchy: y>
a                   1024    True
b                   1024    True
<<U1>               <int64> <bool>
>>> ih1.difference(ih2)
<IndexHierarchy>
a                2048    False
a                2048    True
<<U1>            <int64> <bool>

IndexHierarchy.dropfalsy(*, condition=<function all>)[source]

Return a new obj:IndexHierarchy after removing labels where any or all values are falsy. The condition is determined by a NumPy ufunc that process the Boolean array returned by isna(); the default is np.all.

Parameters:

* –
condition –

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, True), ('', 0, False), ('b', 1024, True)), name='x')
>>> ih
<IndexHierarchy: x>
a                   1024    True
                    0       False
b                   1024    True
<<U1>               <int64> <bool>
>>> ih.dropfalsy()
<IndexHierarchy: x>
a                   1024    True
b                   1024    True
<<U1>               <int64> <bool>

IndexHierarchy.dropna(*, condition=<function all>)[source]

Return a new obj:IndexHierarchy after removing labels where any or all values are NA (NaN or None). The condition is determined by a NumPy ufunc that process the Boolean array returned by isna(); the default is np.all.

Parameters:

* –
condition –

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024), (1, 2048), (np.nan, np.nan)), name=('x', 'y'))
>>> ih
<IndexHierarchy: ('x', 'y')>
0.0                          1024.0
1.0                          2048.0
nan                          nan
<float64>                    <float64>
>>> ih.dropna()
<IndexHierarchy: ('x', 'y')>
0.0                          1024.0
1.0                          2048.0
<float64>                    <float64>

IndexHierarchy.equals(other, *, compare_name=False, compare_dtype=False, compare_class=False, skipna=True)[source]

Return a bool from comparison to any other object.

Parameters:

compare_name – Include equality of the container’s name (and all composed containers) in the comparison.
compare_dtype – Include equality of the container’s dtype (and all composed containers) in the comparison.
compare_class – Include equality of the container’s class (and all composed containers) in the comparison.
skipna – If True, comparisons between missing values are equal.

>>> ih1 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih1
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih2 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('b', 1024, True)), name='y')
>>> ih2
<IndexHierarchy: y>
a                   1024    True
b                   1024    True
<<U1>               <int64> <bool>
>>> ih1.equals(ih2)
False

IndexHierarchy.fillfalsy(value)[source]

Return an IndexHierarchy after replacing falsy values with the supplied value.

Parameters:: value – Value to be used to replace missing values (NaN or None).

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, True), ('', 0, False), ('b', 1024, True)), name='x')
>>> ih
<IndexHierarchy: x>
a                   1024    True
                    0       False
b                   1024    True
<<U1>               <int64> <bool>
>>> ih.fillfalsy(-1)
<IndexHierarchy: x>
a                   1024    True
-1                  -1      -1
b                   1024    True
<object>            <int64> <object>

IndexHierarchy.fillna(value)[source]

Return an IndexHierarchy after replacing NA (NaN or None) with the supplied value.

Parameters:: value – Value to be used to replace missing values (NaN or None).

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024), (1, 2048), (np.nan, np.nan)), name=('x', 'y'))
>>> ih
<IndexHierarchy: ('x', 'y')>
0.0                          1024.0
1.0                          2048.0
nan                          nan
<float64>                    <float64>
>>> ih.fillna(0)
<IndexHierarchy: ('x', 'y')>
0.0                          1024.0
1.0                          2048.0
0.0                          0.0
<float64>                    <float64>

IndexHierarchy.flat()[source]

Return a flat, one-dimensional index of tuples for each level.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, '1517-04-01'), ('a', 2048, '1789-12-31'), ('b', 0, '1620-11-21')), index_constructors=(sf.Index, sf.Index, sf.IndexDate), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
a                                 1024    1517-04-01
a                                 2048    1789-12-31
b                                 0       1620-11-21
<<U1>                             <int64> <datetime64[D]>
>>> ih.flat().display_wide()
<Index: ('x', 'y', 'z')>
('a', 1024, numpy.datetime64('1517-04-01'))
('a', 2048, numpy.datetime64('1789-12-31'))
('b', 0, numpy.datetime64('1620-11-21'))
<object>

IndexHierarchy.head(count=5)

Return a Index consisting only of the top elements as specified by count.

Parameters:: count – Number of elements to be returned from the top of the Index

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.head(2)
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
<<U1>               <datetime64[D]>

IndexHierarchy.iloc_searchsorted(values, *, side_left=True)[source]

Given a sorted Series, return the iloc (integer) position(s) at which insertion in values would retain sort order.

Parameters:

values – a single value, or iterable of values.
side_left – If True, the index of the first suitable location found is given, else return the last such index. If matching an existing value, side_left==True will return that position, side_left==Right will return the next position (or the length).

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.iloc_searchsorted(('b', np.datetime64('1620-11-21')))
3

IndexHierarchy.index_at_depth(depth_level=0)[source]

Return an index, or a tuple of indices for the depth_level specified.

Parameters:: depth_level – a single depth level, or iterable depth of depth levels.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, '1517-04-01'), ('a', 2048, '1789-12-31'), ('b', 0, '1620-11-21')), index_constructors=(sf.Index, sf.Index, sf.IndexDate), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
a                                 1024    1517-04-01
a                                 2048    1789-12-31
b                                 0       1620-11-21
<<U1>                             <int64> <datetime64[D]>
>>> ih.index_at_depth(0)
<Index>
a
b
<<U1>
>>> ih.index_at_depth([2, 0])
(<IndexDate>
1517-04-01
1789-12-31
1620-11-21
<datetime64[D]>, <Index>
a
b
<<U1>)

IndexHierarchy.indexer_at_depth(depth_level=0)[source]

Return the indexers for the depth_level specified. Array will 2D if multiple depths are selected.

Parameters:: depth_level – a single depth level, or iterable depth of depth levels.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, '1517-04-01'), ('a', 2048, '1789-12-31'), ('b', 0, '1620-11-21')), index_constructors=(sf.Index, sf.Index, sf.IndexDate), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
a                                 1024    1517-04-01
a                                 2048    1789-12-31
b                                 0       1620-11-21
<<U1>                             <int64> <datetime64[D]>
>>> ih.indexer_at_depth(0)
[0 0 1]
>>> ih.indexer_at_depth([2, 0])
[[0 1 2]
 [0 0 1]]

IndexHierarchy.intersection(*others)[source]

Perform intersection with one or many Index, container, or NumPy array. Identical comparisons retain order.

>>> ih1 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih1
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih2 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('b', 1024, True)), name='y')
>>> ih2
<IndexHierarchy: y>
a                   1024    True
b                   1024    True
<<U1>               <int64> <bool>
>>> ih1.intersection(ih2)
<IndexHierarchy>
a                1024    True
<<U1>            <int64> <bool>

IndexHierarchy.isfalsy()

Return a same-shaped, Boolean ndarray indicating which values are falsy.

>>> ih = sf.IndexHierarchy.from_labels((('b', np.nan, True), ('', 4096, None), ('c', 0, False)), name='y')
>>> ih.isfalsy()
[[False  True False]
 [ True False  True]
 [False  True  True]]

IndexHierarchy.isin(other)[source]

Return a Boolean array showing True where one or more of the passed in iterable of labels is found in the index.

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.isin((('c', np.datetime64('1517-04-01')), ('a', np.datetime64('1620-11-21'))))
[False  True False False  True False]

IndexHierarchy.isna()

Return a same-shaped, Boolean ndarray indicating which values are NaN or None.

>>> ih = sf.IndexHierarchy.from_labels((('b', np.nan, True), ('', 4096, None), ('c', 0, False)), name='y')
>>> ih.isna()
[[False  True False]
 [False False  True]
 [False False False]]

IndexHierarchy.label_widths_at_depth(depth_level=0)[source]

A generator of pairs, where each pair is the label and the contiguous count of that label found at the depth specified by depth_level.

Parameters:: depth_level – a depth level, starting from zero.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False), ('b', 1024, True)), name='x')
>>> ih
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
b                   1024    True
<<U1>               <int64> <bool>
>>> tuple(ih.label_widths_at_depth(0))
(('a', 3), ('b', 1))
>>> tuple(ih.label_widths_at_depth(1))
((1024, 1), (2048, 2), (1024, 1))
>>> tuple(ih.label_widths_at_depth(2))
((True, 2), (False, 1), (True, 1))

IndexHierarchy.level_add(level, *, index_constructor=None)[source]

Return an IndexHierarchy with a new root (outer) level added.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024), (1, 2048), (np.nan, np.nan)), name=('x', 'y'))
>>> ih
<IndexHierarchy: ('x', 'y')>
0.0                          1024.0
1.0                          2048.0
nan                          nan
<float64>                    <float64>
>>> ih.level_add('A')
<IndexHierarchy: ('x', 'y')>
A                            0.0       1024.0
A                            1.0       2048.0
A                            nan       nan
<<U1>                        <float64> <float64>

IndexHierarchy.level_drop(count=1)[source]

Return an IndexHierarchy with one or more leaf levels removed.

Parameters:: count – A positive value is the number of depths to remove from the root (outer) side of the hierarchy; a negative value is the number of depths to remove from the leaf (inner) side of the hierarchy.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, '1517-04-01'), ('a', 2048, '1789-12-31'), ('b', 0, '1620-11-21')), index_constructors=(sf.Index, sf.Index, sf.IndexDate), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
a                                 1024    1517-04-01
a                                 2048    1789-12-31
b                                 0       1620-11-21
<<U1>                             <int64> <datetime64[D]>
>>> ih.level_drop()
<IndexHierarchy: ('y', 'z')>
1024                         1517-04-01
2048                         1789-12-31
0                            1620-11-21
<int64>                      <datetime64[D]>

IndexHierarchy.loc_searchsorted(values, *, side_left=True, fill_value=nan)[source]

Given a sorted Series, return the loc (label) position(s) at which insertion in values would retain sort order.

Parameters:

values – a single value, or iterable of values.
side_left – If True, the index of the first suitable location found is given, else return the last such index. If matching an existing value, side_left==True will return that position, side_left==Right will return the next position (or the length).
fill_value – A value to be used to fill the label beyond the last label.

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.loc_searchsorted(('b', np.datetime64('1620-11-21')))
('b', numpy.datetime64('1620-11-21'))

IndexHierarchy.loc_to_iloc(key)[source]

Given a label (loc) style key (either a label, a list of labels, a slice, an HLoc object, or a Boolean selection), return the index position (iloc) style key. Keys that are not found will raise a KeyError or a sf.LocInvalid error.

Parameters:: key – a label key.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False), ('b', 1024, True)), name='x')
>>> ih
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
b                   1024    True
<<U1>               <int64> <bool>
>>> ih.loc_to_iloc(('b', 1024, True))
3
>>> ih.loc_to_iloc([('a', 1024, True), ('b', 1024, True)])
[0, 3]
>>> ih.loc_to_iloc(slice(('a', 2048, False), None))
slice(2, None, None)

IndexHierarchy.max(axis=0, skipna=True)

Return the maximum along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.max()
[   1 1024   32]

IndexHierarchy.mean(axis=0, skipna=True, out=None)

Return the mean along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.mean()
NotImplementedError('<function mean at 0x7f65f318ab60> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.median(axis=0, skipna=True, out=None)

Return the median along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.median()
NotImplementedError('<function median at 0x7f65caa8ade0> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.min(axis=0, skipna=True, out=None)

Return the minimum along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.min()
[   0 1024   32]

IndexHierarchy.notfalsy()

Return a same-shaped, Boolean ndarray indicating which values are falsy.

>>> ih = sf.IndexHierarchy.from_labels((('b', np.nan, True), ('', 4096, None), ('c', 0, False)), name='y')
>>> ih.notfalsy()
[[ True False  True]
 [False  True False]
 [ True False False]]

IndexHierarchy.notna()

Return a same-shaped, Boolean ndarray indicating which values are NaN or None.

>>> ih = sf.IndexHierarchy.from_labels((('b', np.nan, True), ('', 4096, None), ('c', 0, False)), name='y')
>>> ih.notna()
[[ True False  True]
 [ True  True False]
 [ True  True  True]]

IndexHierarchy.prod(axis=0, skipna=True, allna=1, out=None)

Return the product along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.prod()
NotImplementedError('<function prod at 0x7f65f318a480> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.rehierarch(depth_map, *, index_constructors=None)[source]

Return a new IndexHierarchy that conforms to the new depth assignments given be depth_map.

>>> ih = sf.IndexHierarchy.from_labels((('b', 1024, True), ('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih
<IndexHierarchy: x>
b                   1024    True
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih.rehierarch([2, 0, 1])
<IndexHierarchy>
True             b     1024
True             a     1024
True             a     2048
False            a     2048
<bool>           <<U1> <int64>

IndexHierarchy.relabel(mapper)[source]

Return a new IndexHierarchy with labels replaced by the callable or mapping; order will be retained. If a mapping is used, the mapping should map tuple representation of labels, and need not map all origin keys.

>>> ih = sf.IndexHierarchy.from_labels((('b', 1024, True), ('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih
<IndexHierarchy: x>
b                   1024    True
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih.relabel(lambda l: (l[0], l[1], str(l[2])[0]))
<IndexHierarchy: x>
b                   1024    T
a                   1024    T
a                   2048    T
a                   2048    F
<<U1>               <int64> <<U1>

IndexHierarchy.relabel_at_depth(mapper, depth_level=0)[source]

Return a new IndexHierarchy after applying mapper to a level or each individual level specified by depth_level.

mapper can be a callable, mapping, or iterable.

If a callable, it must accept a single hashable, and return a single hashable.
If a mapping, it must map a single hashable to a single hashable.
If a iterable, it must be the same length as self.

This call:

>>> index.relabel_at_depth(mapper, depth_level=[0, 1, 2])

is equivalent to:

>>> for level in [0, 1, 2]:
>>>     index = index.relabel_at_depth(mapper, depth_level=level)

albeit more efficient.

>>> ih = sf.IndexHierarchy.from_labels((('b', 1024, True), ('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih
<IndexHierarchy: x>
b                   1024    True
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih.relabel_at_depth(lambda l: str(l)[0], depth_level=[1, 2])
<IndexHierarchy: x>
b                   1     T
a                   1     T
a                   2     T
a                   2     F
<<U1>               <<U1> <<U1>

IndexHierarchy.rename(name)[source]

Return a new IndexHierarchy with an updated name attribute.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False), ('b', 1024, True)), name='x')
>>> ih
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
b                   1024    True
<<U1>               <int64> <bool>
>>> ih.rename('y')
<IndexHierarchy: y>
a                   1024    True
a                   2048    True
a                   2048    False
b                   1024    True
<<U1>               <int64> <bool>

IndexHierarchy.roll(shift)[source]

Return an IndexHierarchy with values rotated forward and wrapped around (with a positive shift) or backward and wrapped around (with a negative shift).

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.roll(2)
<IndexHierarchy: x>
c                   1517-04-01
c                   1620-11-21
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
<<U1>               <datetime64[D]>

IndexHierarchy.sample(count=1, *, seed=None)

Randomly (optionally made deterministic with a fixed seed) extract items from the container to return a subset of the container.

Parameters:

select. (Number of elements to) –
selection. (Initial state of random) –

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.sample(2, seed=0)
<IndexHierarchy: x>
b                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>

IndexHierarchy.sort(*, ascending=True, kind='mergesort', key=None)[source]

Return a new Index with the labels sorted.

Parameters:

ascendings – Boolean, or iterable of Booleans; if True, the lowest ranks correspond to the lowest values; if an iterable, apply per column or row. The default is True.
kind – Name of the sort algorithm as passed to NumPy.
key – A function that is used to pre-process the selected columns or rows and derive new values to sort by.

>>> ih = sf.IndexHierarchy.from_labels((('b', 1024, True), ('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih
<IndexHierarchy: x>
b                   1024    True
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih.sort()
<IndexHierarchy: x>
a                   1024    True
a                   2048    False
a                   2048    True
b                   1024    True
<<U1>               <int64> <bool>
>>> ih.sort(ascending=False)
<IndexHierarchy: x>
b                   1024    True
a                   2048    True
a                   2048    False
a                   1024    True
<<U1>               <int64> <bool>

IndexHierarchy.std(axis=0, skipna=True, ddof=0, out=None)

Return the standard deviaton along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.std()
NotImplementedError('functools.partial(<function std at 0x7f65f3190c30>, ddof=0) for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.sum(axis=0, skipna=True, allna=0, out=None)

Sum values along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.sum()
NotImplementedError('<function sum at 0x7f65f3189a80> for IndexHierarchy is not defined; convert to `Frame`.')

IndexHierarchy.tail(count=5)

Return a Index consisting only of the bottom elements as specified by count.

Parameters:: count – Number of elements to be returned from the bottom of the Index

>>> ih = sf.IndexHierarchy.from_product(('a', 'b', 'c'), ('1517-04-01', '1620-11-21'), name='x', index_constructors=(sf.Index, sf.IndexDate))
>>> ih
<IndexHierarchy: x>
a                   1517-04-01
a                   1620-11-21
b                   1517-04-01
b                   1620-11-21
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>
>>> ih.tail(2)
<IndexHierarchy: x>
c                   1517-04-01
c                   1620-11-21
<<U1>               <datetime64[D]>

IndexHierarchy.union(*others)[source]

Perform union with another Index, container, or NumPy array. Identical comparisons retain order.

>>> ih1 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('a', 2048, True), ('a', 2048, False)), name='x')
>>> ih1
<IndexHierarchy: x>
a                   1024    True
a                   2048    True
a                   2048    False
<<U1>               <int64> <bool>
>>> ih2 = sf.IndexHierarchy.from_labels((('a', 1024, True), ('b', 1024, True)), name='y')
>>> ih2
<IndexHierarchy: y>
a                   1024    True
b                   1024    True
<<U1>               <int64> <bool>
>>> ih1.union(ih2)
<IndexHierarchy>
a                2048    False
a                1024    True
b                1024    True
a                2048    True
<<U1>            <int64> <bool>

IndexHierarchy.unique(depth_level=0, order_by_occurrence=False)[source]

Return a NumPy array of unique values.

Parameters:

depth_level – Specify a single depth or multiple depths in an iterable.
order_by_occurrence – if True, values are ordered by when they first appear

Returns:

numpy.ndarray

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024), (1, 2048), (np.nan, np.nan)), name=('x', 'y'))
>>> ih
<IndexHierarchy: ('x', 'y')>
0.0                          1024.0
1.0                          2048.0
nan                          nan
<float64>                    <float64>
>>> ih.unique()
[ 0.  1. nan]

IndexHierarchy.values_at_depth(depth_level=0)[source]

Return an NP array for the depth_level specified.

Parameters:: depth_level – a single depth level, or iterable depth of depth levels.

>>> ih = sf.IndexHierarchy.from_labels((('a', 1024, '1517-04-01'), ('a', 2048, '1789-12-31'), ('b', 0, '1620-11-21')), index_constructors=(sf.Index, sf.Index, sf.IndexDate), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
a                                 1024    1517-04-01
a                                 2048    1789-12-31
b                                 0       1620-11-21
<<U1>                             <int64> <datetime64[D]>
>>> ih.values_at_depth(0)
['a' 'a' 'b']
>>> ih.values_at_depth(2)
['1517-04-01' '1789-12-31' '1620-11-21']

IndexHierarchy.var(axis=0, skipna=True, ddof=0, out=None)

Return the variance along the specified axis.

Parameters:

axis – Axis, defaulting to axis 0.
skipna – Skip missing (NaN) values, defaulting to True.

>>> ih = sf.IndexHierarchy.from_labels(((0, 1024, 32), (1, -2048, 32), (1, 1024, 32)), name=('x', 'y', 'z'))
>>> ih
<IndexHierarchy: ('x', 'y', 'z')>
0                                 1024    32
1                                 -2048   32
1                                 1024    32
<int64>                           <int64> <int64>
>>> ih.var()
NotImplementedError('functools.partial(<function var at 0x7f65f3190e30>, ddof=0) for IndexHierarchy is not defined; convert to `Frame`.')