describe.pyc
1 o 2 c��c�3 � @ s` d Z ddlmZ ddlmZmZ ddlmZmZm Z m 3 Z 4 m Z m Z ddl Z ddlZddlmZ ddlmZmZmZ ddlmZ dd lmZ dd 5 lmZm Z m Z m Z m Z m Z ddl!Z"dd l#m$Z$ dd l%m&Z& erndd l!m'Z'm(Z( d8dd�Z)G dd� de�Z*G d d � d e*�Z+G d d � d e*�Z,d9d#d$�Z-d:d(d)�Z.d;d,d-�Z/d;d.d/�Z0d<d0d1�Z1d=d3d4�Z2d>d6d7�Z3dS )?z� 6 Module responsible for execution of NDFrame.describe() method. 7 8 Method NDFrame.describe() delegates actual execution to function describe_ndframe(). 9 � )� annotations)�ABC�abstractmethod)� TYPE_CHECKING�Any�Callable�Hashable�Sequence�castN)� Timestamp)�DtypeObj�NDFrameT�npt)�find_stack_level)�validate_percentile)� is_bool_dtype�is_complex_dtype�is_datetime64_any_dtype�is_extension_array_dtype�is_numeric_dtype�is_timedelta64_dtype)�concat)�format_percentiles)� DataFrame�Series�objr �include�str | Sequence[str] | None�exclude�datetime_is_numeric�bool� percentiles�#Sequence[float] | np.ndarray | None�returnc C sR t |�}| jdkrttd| �|d�}n ttd| �|||d�}|j|d�}tt|�S )a� Describe series or dataframe. 10 11 Called from pandas.core.generic.NDFrame.describe() 12 13 Parameters 14 ---------- 15 obj: DataFrame or Series 16 Either dataframe or series to be described. 17 include : 'all', list-like of dtypes or None (default), optional 18 A white list of data types to include in the result. Ignored for ``Series``. 19 exclude : list-like of dtypes or None (default), optional, 20 A black list of data types to omit from the result. Ignored for ``Series``. 21 datetime_is_numeric : bool, default False 22 Whether to treat datetime dtypes as numeric. 23 percentiles : list-like of numbers, optional 24 The percentiles to include in the output. All should fall between 0 and 1. 25 The default is ``[.25, .5, .75]``, which returns the 25th, 50th, and 26 75th percentiles. 27 28 Returns 29 ------- 30 Dataframe or series description. 31 � r �r r r )r r r r )r! )�refine_percentiles�ndim�SeriesDescriberr 32 �DataFrameDescriber�describer )r r r r r! Z describer�result� r, ��C:\Users\Jacks.GUTTSPC\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\LocalCache\local-packages\Python310\site-packages\pandas\core\describe.py�describe_ndframe6 s 33 �� 34 r. c @ s( e Zd ZdZddd �Zedd d ��ZdS )�NDFrameDescriberAbstractz�Abstract class for describing dataframe or series. 35 36 Parameters 37 ---------- 38 obj : Series or DataFrame 39 Object to be described. 40 datetime_is_numeric : bool 41 Whether to treat datetime dtypes as numeric. 42 r �DataFrame | Seriesr r r# �Nonec C s || _ || _d S �Nr% )�selfr r r, r, r- �__init__t s 43 z!NDFrameDescriberAbstract.__init__r! � Sequence[float] | np.ndarrayc C s dS )z�Do describe either series or dataframe. 44 45 Parameters 46 ---------- 47 percentiles : list-like of numbers 48 The percentiles to include in the output. 49 Nr, )r3 r! r, r, r- r* x s z!NDFrameDescriberAbstract.describeN)r r0 r r r# r1 )r! r5 r# r0 )�__name__� 50 __module__� __qualname__�__doc__r4 r r* r, r, r, r- r/ i s 51 52 53 r/ c @ s$ e Zd ZU dZded< d 54 dd�Zd S ) r( z2Class responsible for creating series description.r r r! r5 r# c C s t | j| j�}|| j|�S r2 )�select_describe_funcr r )r3 r! � describe_funcr, r, r- r* � s 55 � zSeriesDescriber.describeN)r! r5 r# r )r6 r7 r8 r9 �__annotations__r* r, r, r, r- r( � s 56 r( c s4 e Zd ZdZd� fd d � Zddd�Zdd� Z� ZS )r) a� Class responsible for creating dataobj description. 57 58 Parameters 59 ---------- 60 obj : DataFrame 61 DataFrame to be described. 62 include : 'all', list-like of dtypes or None 63 A white list of data types to include in the result. 64 exclude : list-like of dtypes or None 65 A black list of data types to omit from the result. 66 datetime_is_numeric : bool 67 Whether to treat datetime dtypes as numeric. 68 r r r r r r r r# r1 c s> || _ || _|jdkr|jjdkrtd��t� j||d� d S )N� r z+Cannot describe a DataFrame without columns)r )r r r' �columns�size� 69 ValueError�superr4 )r3 r r r r �� __class__r, r- r4 � s 70 zDataFrameDescriber.__init__r! r5 c sn | � � }g }|�� D ]\}}t|| j�}|�|||�� q 71 t|�� t� fdd�|D �ddd�}|j�� |_|S )Nc s g | ] }|j � d d��qS )F)�copy)�reindex��.0�x�� col_namesr, r- � 72 <listcomp>� s z/DataFrameDescriber.describe.<locals>.<listcomp>r$ F)�axis�sort) � _select_data�itemsr: r �append�reorder_columnsr r> rD )r3 r! �data�ldesc�_�seriesr; �dr, rI r- r* � s � zDataFrameDescriber.describec C s� | j du r)| jdu r)tjg}| jr|�d� | jj|d�}t|j �dkr'| j}|S | j dkr>| jdur9d}t 73 |��| j}|S | jj| j | jd�}|S )z Select columns to be described.N�datetime)r r �allz*exclude must be None when include is 'all')r r ) r r �np�numberr rP r � select_dtypes�lenr> r@ )r3 Zdefault_includerR �msgr, r, r- rN � s& 74 75 � 76 ��z DataFrameDescriber._select_data) 77 r r r r r r r r r# r1 )r! r5 r# r )r6 r7 r8 r9 r4 r* rN � __classcell__r, r, rB r- r) � s 78 79 r) rS �Sequence[Series]�list[Hashable]c C sD g }t dd� | D �td�}|D ]}|D ] }||vr |�|� qq|S )z,Set a convenient order for rows for display.c s s � | ]}|j V qd S r2 )�indexrF r, r, r- � <genexpr>� s � z"reorder_columns.<locals>.<genexpr>)�key)�sortedr\ rP )rS �namesZ ldesc_indexesZidxnames�namer, r, r- rQ � s 80 ��rQ rU r �Sequence[float]c C s� ddl m} t|�}g d�| dg }| �� | �� | �� | �� g| �|��� | � � g }t 81 | �r6t � � }nt | �rDt| �sDt�d�}nd}|||| j|d�S )z�Describe series containing numerical data. 82 83 Parameters 84 ---------- 85 series : Series 86 Series to be described. 87 percentiles : list-like of numbers 88 The percentiles to include in the output. 89 r �r )�count�mean�std�min�max�floatN�ra rf �dtype)�pandasr r ri rj rk rl �quantile�tolistrm r �pd� Float64Dtyper r rY rp rf )rU r! r �formatted_percentiles� 90 stat_indexrV rp r, r, r- �describe_numeric_1d� s 91 ��� 92 rx rR �percentiles_ignoredc 93 C s� g d�}| � � }t||dk �}|dkr"|jd |jd }}d}n tjtj}}d}| �� |||g}ddlm} | ||| j |d�S )z�Describe series containing categorical data. 94 95 Parameters 96 ---------- 97 data : Series 98 Series to be described. 99 percentiles_ignored : list-like of numbers 100 Ignored, but in place to unify interface. 101 )ri �unique�top�freqr N�objectrh ro ) 102 � value_countsr\ ra �ilocrY �nanri rq r rf ) 103 rR ry re � objcounts� count_uniquer{ r| rp r+ r r, r, r- �describe_categorical_1d s r� c C s 104 ddg}| � � }t||dk �}| �� |g}d}|dkrf|jd |jd }}| jj} | �� j� d�} 105 t 106 |�}|j durF| durF|� | �}n|� | �}|g d�7 }|||t 107 | 108 �� | d�t 109 | 110 �� | d�g7 }n|dd g7 }|tjtjg7 }d 111 }dd lm} | ||| j|d �S ) z�Describe series containing timestamp data treated as categorical. 112 113 Parameters 114 ---------- 115 data : Series 116 Series to be described. 117 percentiles_ignored : list-like of numbers 118 Ignored, but in place to unify interface. 119 ri rz r N�i8)r{ r| �first�last)�tzr{ r| r} rh ro )r~ r\ ri ra r �dtr� �dropna�values�viewr �tzinfo� 120 tz_convert� tz_localizerl rm rY r� rq r rf ) rR ry re r� r� r+ rp r{ r| r� �asintr r, r, r- �$describe_timestamp_as_categorical_1d! s2 121 � 122 r� c C sd ddl m} t|�}g d�| dg }| �� | �� | �� g| �|��� | �� g }|||| j d�S )z�Describe series containing datetime64 dtype. 123 124 Parameters 125 ---------- 126 data : Series 127 Series to be described. 128 percentiles : list-like of numbers 129 The percentiles to include in the output. 130 r rh )ri rj rl rm )ra rf ) 131 rq r r ri rj rl rr rs rm rf )rR r! r rv rw rV r, r, r- �describe_timestamp_1dQ s ���r� r c C sT t | j�rtS t| �r tS t| j�r!|rtS tjdt t 132 � d� t S t | j�r(tS tS )z�Select proper function for describing series based on data type. 133 134 Parameters 135 ---------- 136 data : Series 137 Series to be described. 138 datetime_is_numeric : bool 139 Whether to treat datetime dtypes as numeric. 140 z�Treating datetime data as categorical rather than numeric in `.describe` is deprecated and will be removed in a future version of pandas. Specify `datetime_is_numeric=True` to silence this warning and adopt the future behavior now.)� 141 stacklevel) r rp r� r rx r r� �warnings�warn� FutureWarningr r� r )rR r r, r, r- r: i s 142 143 � 144 r: �%np.ndarray[Any, np.dtype[np.float64]]c C st | du r t �g d��S t| �} t| � d| vr | �d� t �| �} t �| �}| dus,J �t|�t| �k r8td��|S )z� 145 Ensure that percentiles are unique and sorted. 146 147 Parameters 148 ---------- 149 percentiles : list-like of numbers, optional 150 The percentiles to include in the output. 151 N)g �?� �?g �?r� z%percentiles cannot contain duplicates) rY �array�listr rP �asarrayrz r\ r@ )r! � unique_pctsr, r, r- r& � s 152 153 154 r& ) r r r r r r r r r! r" r# r )rS r_ r# r` )rU r r! rg r# r )rR r ry rg r# r )rR r r! rg r# r )rR r r r r# r )r! r" r# r� )4r9 � 155 __future__r �abcr r �typingr r r r r r 156 r� �numpyrY �pandas._libs.tslibsr �pandas._typingr r r �pandas.util._exceptionsr �pandas.util._validatorsr �pandas.core.dtypes.commonr r r r r r rq rt �pandas.core.reshape.concatr �pandas.io.formats.formatr r r r. r/ r( r) rQ rx r� r� r� r: r& r, r, r, r- �<module> s6 157 3 158 G 159 160 161 162 0 163 $