索引和遍历

import pandas as pd
import numpy as np

以字典的方式索引

df = pd.DataFrame(np.random.rand(3, 2), columns=['first', 'second'], index=['a', 'b', 'c'])
df

	first	second
a	0.241914	0.936741
b	0.721130	0.812072
c	0.707245	0.819285

"""
可直接取某列
不能直接取某行，如df['a']会报错
"""
df['first']

a    0.241914
b    0.721130
c    0.707245
Name: first, dtype: float64

# 先column，再index
df['first']['b']

0.7211298502376191

loc与iloc

# 使用loc取某行
df.loc['a']

first     0.241914
second    0.936741
Name: a, dtype: float64

# 使用loc时，先index，再column
df.loc[:'b', :'first']

	first
a	0.241914
b	0.721130

# 使用iloc时，索引方式和数组相同
df.iloc[:2, :2]

	first	second
a	0.241914	0.936741
b	0.721130	0.812072

# 修改数值
df.iloc[2, 1] = 100
df

	first	second
a	0.241914	0.936741
b	0.721130	0.812072
c	0.707245	100.000000

# 直接修改df.values
df.values[0, 1] = 1
df

	first	second
a	0.241914	1.000000
b	0.721130	0.812072
c	0.707245	100.000000

按条件索引

df[df['second'] >= 0.5]

	first	second
a	0.241914	1.000000
b	0.721130	0.812072
c	0.707245	100.000000

df[(df['first'] >= 0.5) & (df['first'] <= 1.0)]

	first	second
b	0.721130	0.812072
c	0.707245	100.000000

遍历dataframe

# 按行遍历
for x, y in df.iterrows():
    # index
    print(x)
    # 此行的series
    print(y)
    print(y['second'])

a
first     0.241914
second    1.000000
Name: a, dtype: float64
1.0
b
first     0.721130
second    0.812072
Name: b, dtype: float64
0.8120721605424106
c
first       0.707245
second    100.000000
Name: c, dtype: float64
100.0

# 按列遍历
for x, y in df.items():
    # column
    print(x)
    # 此列的series
    print(y)

first
a    0.241914
b    0.721130
c    0.707245
Name: first, dtype: float64
second
a      1.000000
b      0.812072
c    100.000000
Name: second, dtype: float64