Python数据分析——pandas

最新推荐文章于 2024-06-28 16:16:19 发布

鸢鸢z

最新推荐文章于 2024-06-28 16:16:19 发布

阅读量320

点赞数 2

分类专栏：数据分析文章标签： python 数据分析

本文链接：https://blog.csdn.net/koori_145/article/details/107850127

版权

数据分析专栏收录该内容

3 篇文章 0 订阅

订阅专栏

（本专栏是我的慕课学习笔记，后续发现不足之处会更新）

pandas基础操作：

>>> import numpy as np
>>> import pandas as pd

>>> dates = pd.date_range("20200806", periods=8)    # 连续的8天
>>> # randn为标准正态分布，index为key值，columns为属性值
>>> df = pd.DataFrame(np.random.randn(8, 5), index=dates, columns=list("ABCDE"))

>>> print("整个表：\n", df)                       # 打印整个表
                    A         B         C         D         E
2020-08-06  0.707288  0.499962  0.130611  1.764557 -1.165358
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789
2020-08-09 -0.012218 -1.224213 -1.319811  0.323090  0.063972
2020-08-10  0.297825  0.546069 -0.192028  0.413679 -1.113024
2020-08-11 -1.070570 -0.530023 -0.301933  0.538606 -1.231873
2020-08-12 -0.572311 -2.700809  0.258386  0.114140  0.487276
2020-08-13  0.050315  0.475426  0.581819 -2.242387  1.220767


>>> print("\n表的前3行：\n", df.head(3))          # 打印表的前3行
                    A         B         C         D         E
2020-08-06  0.707288  0.499962  0.130611  1.764557 -1.165358
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789


>>> print("\n表的后3行：\n", df.tail(3))          # 打印表的后3行
                    A         B         C         D         E
2020-08-11 -1.070570 -0.530023 -0.301933  0.538606 -1.231873
2020-08-12 -0.572311 -2.700809  0.258386  0.114140  0.487276
2020-08-13  0.050315  0.475426  0.581819 -2.242387  1.220767


>>> print("\nindex值：\n", df.index)              # 只打印index值
index值：
 DatetimeIndex(['2020-08-06', '2020-08-07', '2020-08-08', '2020-08-09',
               '2020-08-10', '2020-08-11', '2020-08-12', '2020-08-13'],
              dtype='datetime64[ns]', freq='D')

>>> print("\nvalues值：\n", df.values)            # 只打印values值
 [[ 0.70728844  0.4999616   0.13061059  1.76455701 -1.1653581 ]
 [-0.99735496  1.22000868 -0.45235699 -0.43247666  1.31265841]
 [-0.52401434  0.37303048 -0.01251882 -0.05932329 -2.15578928]
 [-0.01221819 -1.22421337 -1.31981069  0.32308968  0.06397238]
 [ 0.29782485  0.54606854 -0.19202779  0.41367944 -1.11302353]
 [-1.07056955 -0.53002278 -0.30193276  0.53860645 -1.23187343]
 [-0.57231052 -2.70080883  0.2583857   0.11414044  0.48727611]
 [ 0.05031547  0.47542649  0.58181917 -2.24238729  1.22076725]]


>>> print("\n转置：\n", df.T)                     # 转置
    2020-08-06  2020-08-07  2020-08-08  ...  2020-08-11  2020-08-12  2020-08-13
A    0.707288   -0.997355   -0.524014  ...   -1.070570   -0.572311    0.050315
B    0.499962    1.220009    0.373030  ...   -0.530023   -2.700809    0.475426
C    0.130611   -0.452357   -0.012519  ...   -0.301933    0.258386    0.581819
D    1.764557   -0.432477   -0.059323  ...    0.538606    0.114140   -2.242387
E   -1.165358    1.312658   -2.155789  ...   -1.231873    0.487276    1.220767

[5 rows x 8 columns]


>>> print("\n按A属性排序：\n", df.sort_values("A"))    # 按A属性递增排序
# axis=0时，按index倒序输出；axis=1时，属性从"ABCDE"变成"EDCBA"
                    A         B         C         D         E
2020-08-11 -1.070570 -0.530023 -0.301933  0.538606 -1.231873
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658
2020-08-12 -0.572311 -2.700809  0.258386  0.114140  0.487276
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789
2020-08-09 -0.012218 -1.224213 -1.319811  0.323090  0.063972
2020-08-13  0.050315  0.475426  0.581819 -2.242387  1.220767
2020-08-10  0.297825  0.546069 -0.192028  0.413679 -1.113024
2020-08-06  0.707288  0.499962  0.130611  1.764557 -1.165358


>>> print("\n对index排序：\n", df.sort_index(axis=0, ascending=False))
                    A         B         C         D         E
2020-08-13  0.050315  0.475426  0.581819 -2.242387  1.220767
2020-08-12 -0.572311 -2.700809  0.258386  0.114140  0.487276
2020-08-11 -1.070570 -0.530023 -0.301933  0.538606 -1.231873
2020-08-10  0.297825  0.546069 -0.192028  0.413679 -1.113024
2020-08-09 -0.012218 -1.224213 -1.319811  0.323090  0.063972
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658
2020-08-06  0.707288  0.499962  0.130611  1.764557 -1.165358


>>> # 所有属性值的数量，平均值，标准差，最小值，下四分位数，中位数，上四分位数，最大值
>>> print("\n描述：\n", df.describe())
               A         B         C         D         E
count  8.000000  8.000000  8.000000  8.000000  8.000000
mean  -0.265130 -0.167569 -0.163479  0.052486 -0.322671
std    0.629314  1.267794  0.571121  1.126887  1.275246
min   -1.070570 -2.700809 -1.319811 -2.242387 -2.155789
25%   -0.678572 -0.703570 -0.339539 -0.152612 -1.181987
50%   -0.268116  0.424228 -0.102273  0.218615 -0.524526
75%    0.112193  0.511488  0.162554  0.444911  0.670649
max    0.707288  1.220009  0.581819  1.764557  1.312658


>>> print("\n只打印A属性：\n", df["A"])            # 只打印A属性
2020-08-06    0.707288
2020-08-07   -0.997355
2020-08-08   -0.524014
2020-08-09   -0.012218
2020-08-10    0.297825
2020-08-11   -1.070570
2020-08-12   -0.572311
2020-08-13    0.050315
Freq: D, Name: A, dtype: float64


>>> print("\n使用切片方式打印前3行：\n", df[:3])   # 打印前3行
                    A         B         C         D         E
2020-08-06  0.707288  0.499962  0.130611  1.764557 -1.165358
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789


>>> print("\n打印指定index区间：\n", df["20200808":"20200811"])    # 打印指定index区间
                    A         B         C         D         E
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789
2020-08-09 -0.012218 -1.224213 -1.319811  0.323090  0.063972
2020-08-10  0.297825  0.546069 -0.192028  0.413679 -1.113024
2020-08-11 -1.070570 -0.530023 -0.301933  0.538606 -1.231873


>>> print("\n打印第1行数据：\n", df.loc[dates[0]])     # 打印第1行数据
A    0.707288
B    0.499962
C    0.130611
D    1.764557
E   -1.165358
Name: 2020-08-06 00:00:00, dtype: float64


>>> print("\n打印指定index指定属性：\n", df.loc["20200808":"20200811", ["A", "C"]])      # 打印指定index指定属性
                    A         C
2020-08-08 -0.524014 -0.012519
2020-08-09 -0.012218 -1.319811
2020-08-10  0.297825 -0.192028
2020-08-11 -1.070570 -0.301933


>>> print("\n打印指定index指定属性的值：\n", df.at[dates[0], "B"])     # 打印指定index指定属性的值
0.4999616023174961


>>> print("\n通过下标打印指定index指定属性：\n", df.iloc[1:4, 2:5])      # 第2~4个index的第3~5列属性
                    C         D         E
2020-08-07 -0.452357 -0.432477  1.312658
2020-08-08 -0.012519 -0.059323 -2.155789
2020-08-09 -1.319811  0.323090  0.063972


>>> print("\n设定条件打印指定值：\n", df[(df.B > 0) & (df.C < 0)])            # 设定条件打印指定值
                    A         B         C         D         E
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789
2020-08-10  0.297825  0.546069 -0.192028  0.413679 -1.113024


>>> print("\n设定条件值打印总表（不满足条件为NaN）：\n", df[df > 0])      # 设定条件值打印总表（不满足条件为NaN）
                    A         B         C         D         E
2020-08-06  0.707288  0.499962  0.130611  1.764557       NaN
2020-08-07       NaN  1.220009       NaN       NaN  1.312658
2020-08-08       NaN  0.373030       NaN       NaN       NaN
2020-08-09       NaN       NaN       NaN  0.323090  0.063972
2020-08-10  0.297825  0.546069       NaN  0.413679       NaN
2020-08-11       NaN       NaN       NaN  0.538606       NaN
2020-08-12       NaN       NaN  0.258386  0.114140  0.487276
2020-08-13  0.050315  0.475426  0.581819       NaN  1.220767


>>> s1 = pd.Series(list(range(101, 109)), index=pd.date_range("20200806", periods=8))
>>> df["F"] = s1        # 添加一列"F"，属性为101~108

>>> print("\n添加一列"F"，属性为101~108：\n", df)
                    A         B         C         D         E    F
2020-08-06  0.707288  0.499962  0.130611  1.764557 -1.165358  101
2020-08-07 -0.997355  1.220009 -0.452357 -0.432477  1.312658  102
2020-08-08 -0.524014  0.373030 -0.012519 -0.059323 -2.155789  103
2020-08-09 -0.012218 -1.224213 -1.319811  0.323090  0.063972  104
2020-08-10  0.297825  0.546069 -0.192028  0.413679 -1.113024  105
2020-08-11 -1.070570 -0.530023 -0.301933  0.538606 -1.231873  106
2020-08-12 -0.572311 -2.700809  0.258386  0.114140  0.487276  107
2020-08-13  0.050315  0.475426  0.581819 -2.242387  1.220767  108


>>> df[df < 0] = -df    # 将表格中所有的负数改为绝对值

>>> print("\n将表格中所有的负数改为绝对值：\n", df)
                    A         B         C         D         E    F
2020-08-06  0.707288  0.499962  0.130611  1.764557  1.165358  101
2020-08-07  0.997355  1.220009  0.452357  0.432477  1.312658  102
2020-08-08  0.524014  0.373030  0.012519  0.059323  2.155789  103
2020-08-09  0.012218  1.224213  1.319811  0.323090  0.063972  104
2020-08-10  0.297825  0.546069  0.192028  0.413679  1.113024  105
2020-08-11  1.070570  0.530023  0.301933  0.538606  1.231873  106
2020-08-12  0.572311  2.700809  0.258386  0.114140  0.487276  107
2020-08-13  0.050315  0.475426  0.581819  2.242387  1.220767  108

缺失值处理：

>>> import numpy as np
>>> import pandas as pd

>>> dates = pd.date_range("20200806", periods=8)    # 连续的8天
>>> # randn为标准正态分布，index为key值，columns为属性值
>>> df = pd.DataFrame(np.random.randn(8, 5), index=dates, columns=list("ABCDE"))

>>> print("整个表：\n", df)         # 打印整个表
                    A         B         C         D         E
2020-08-06  1.480559  2.011120 -0.446669  0.660951 -0.206286
2020-08-07 -0.776498  1.367207 -1.829035  1.205739 -1.435127
2020-08-08 -0.313502 -0.639334  0.623946  0.238073 -1.537367
2020-08-09  0.756416  0.748446  0.561026 -0.651465  0.567481
2020-08-10  0.472599 -1.200840 -1.622537  0.198806  0.991217
2020-08-11 -0.179116  0.477361  0.019794 -0.522681  0.750547
2020-08-12  0.833937  0.567930  0.427194  0.592522  1.281558
2020-08-13  0.797320 -0.833195  0.389737  0.516310  0.276850


>>> df1 = df.reindex(index=dates[:5], columns=list("AC") + ["G"])   # 5行数据，有ACG三个属性
>>> df1.loc[dates[0:3], "G"] = 1234    # 给前3行的G属性赋值为1234

>>> print("整个表：\n", df1)         # 打印整个表
                    A         C       G
2020-08-06  1.480559 -0.446669  1234.0
2020-08-07 -0.776498 -1.829035  1234.0
2020-08-08 -0.313502  0.623946  1234.0
2020-08-09  0.756416  0.561026     NaN
2020-08-10  0.472599 -1.622537     NaN


>>> print("丢弃有空值的行：\n", df1.dropna())       # 丢弃有空值的行
                    A         C       G
2020-08-06  1.480559 -0.446669  1234.0
2020-08-07 -0.776498 -1.829035  1234.0
2020-08-08 -0.313502  0.623946  1234.0


>>> print("给空值赋指定值：\n", df1.fillna(value=9999))       # 丢弃有空值的行
                    A         C       G
2020-08-06  1.480559 -0.446669  1234.0
2020-08-07 -0.776498 -1.829035  1234.0
2020-08-08 -0.313502  0.623946  1234.0
2020-08-09  0.756416  0.561026  9999.0
2020-08-10  0.472599 -1.622537  9999.0

表格数据分析：

>>> import numpy as np
>>> import pandas as pd

>>> dates = pd.date_range("20200806", periods=5)    # 连续的8天
>>> # randn为标准正态分布，index为key值，columns为属性值
>>> df = pd.DataFrame(np.random.randn(5, 3), index=dates, columns=list("ABC"))

>>> print("整个表：\n", df)         # 打印整个表
                    A         B         C
2020-08-06 -0.231186 -0.971756  1.702146
2020-08-07 -0.431149 -0.025855 -0.887768
2020-08-08 -0.693589 -0.253785  1.686983
2020-08-09 -1.332854 -1.194220  0.606153
2020-08-10  0.580246 -0.080983 -0.154577


>>> print("\n各属性平均值：\n", df.mean())     # 各属性平均值
A   -0.421706
B   -0.505320
C    0.590587
dtype: float64


>>> print("\n各属性方差：\n", df.var())     # 各属性方差
A    0.486102
B    0.291341
C    1.294673
dtype: float64


>>> s = pd.Series([1, 3, 5, np.nan, 7], index=dates)
>>> print("\n创建新表格：\n", s)      # 创建新表格
2020-08-06    1.0
2020-08-07    3.0
2020-08-08    5.0
2020-08-09    NaN
2020-08-10    7.0
Freq: D, dtype: float64


>>> print("\n值后移2位，前2位为空，后2位抛弃\n", s.shift(2))      # 值后移2位，前2位为空，后2位抛弃
2020-08-06    NaN
2020-08-07    NaN
2020-08-08    1.0
2020-08-09    3.0
2020-08-10    5.0
Freq: D, dtype: float64


>>> print("\n每个值出现的次数：\n", s.value_counts())            # 每个值出现的次数
7.0    1
5.0    1
3.0    1
1.0    1
dtype: int64


>>> print("\n每个属性的累加值：\n", df.apply(np.cumsum))         # 每个属性的累加值
                    A         B         C
2020-08-06 -0.231186 -0.971756  1.702146
2020-08-07 -0.662336 -0.997611  0.814378
2020-08-08 -1.355924 -1.251396  2.501361
2020-08-09 -2.688778 -2.445616  3.107514
2020-08-10 -2.108532 -2.526599  2.952937


>>> print("\n每个属性的极差值：\n", df.apply(lambda x: x.max() - x.min()))       # 每个属性的极差值
A    1.913100
B    1.168365
C    2.589914
dtype: float64

氷鸢鸢鸢
2020.8.6

鸢鸢z

关注

2
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
Python数据分析——pandas

（本专栏是我的慕课学习笔记，后续发现不足之处会更新）pandas基础操作（无输出版）（有输出的在下面）：import numpy as npimport pandas as pddates = pd.date_range("20200806", periods=8) # 连续的8天# randn为标准正态分布，index为key值，columns为属性值df = pd.DataFrame(np.random.randn(8, 5), index=dates, columns=list("
复制链接

扫一扫