dataframe 的常见操作

import pandas as pd

df0 = pd.DataFrame({"x2": [
                   11.2, 23.2, 23.4, 1], "x3": [11.2, 23.2, 23.4, 12]})

print(df0)

df1 = pd.DataFrame({'x1': ["a", "b", "c", "a"], "x2": [
                   "11.2", "23.2", "23.4", "1"], "x3": ["11.2", "23.2", "23.4", "12"]})
print(df1)
df2 = pd.DataFrame({'x1': ["a", "b", "d"], "x2": ["20.23", 'NaN', "90.4"]})
print(df2)

data = {'brand_model': ['德力西 GW912200', '埃帝尔 IB-207TPN-06231C02', '德力西 72LT3YFM40D', '科顺 2-3654-561 铆钉型'],
        'cate': ['低压', '管阀', '低压', '搬运']}
df3 = pd.DataFrame(data)
print(df3)

df5 = pd.DataFrame({'x1': ["12", "23", "99", "123"], "x2": [
                   "11.2", "23.2", "23.4", "1"], "x3": ["11.2", "23.2", "23.4", "12"]})
print(df5)
df6 = pd.DataFrame({'x1': ["12%", "23", "99", "123"], "x2": [
                   "11.2", "23.2%", "23.4", "1"], "x3": ["11.2", "23.2%", "23.4%", "12%"]})

# 判断dataframe数据为空
df7 = pd.DataFrame(data=[], columns=[1,2])
print(df7.empty)

     x2    x3
0  11.2  11.2
1  23.2  23.2
2  23.4  23.4
3   1.0  12.0
  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
3  a     1    12
  x1     x2
0  a  20.23
1  b    NaN
2  d   90.4
              brand_model cate
0            德力西 GW912200   低压
1  埃帝尔 IB-207TPN-06231C02   管阀
2         德力西 72LT3YFM40D   低压
3       科顺 2-3654-561 铆钉型   搬运
    x1    x2    x3
0   12  11.2  11.2
1   23  23.2  23.2
2   99  23.4  23.4
3  123     1    12
True

print(df1)
# df1.loc[:, 'x1'].tolist()[-time]
print(df1.loc[:, 'x1'])
va1 = df1.loc[:, 'x2'].tolist()[-1]
va2 = df1.loc[:, 'x3'].tolist()[-1]

va3 = va1 - va2
print(va3)
print(va2)

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
3  a     1    12
0    a
1    b
2    c
3    a
Name: x1, dtype: object



---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

/var/folders/w6/9k4dzqlj617f06dfby_vk1pr0000gn/T/ipykernel_13699/4084052124.py in <module>
      5 va2 = df1.loc[:, 'x3'].tolist()[-1]
      6 
----> 7 va3 = va1 - va2
      8 print(va3)
      9 print(va2)


TypeError: unsupported operand type(s) for -: 'str' and 'str'

一、索引的最常见操作

1、设置索引

df1.set_index('x1')

# 设置多重索引
print(df1.set_index(['x1', 'x3']))

# drop=False 表示保留原先"YY"列的数据
df1_drop = df1.set_index("x1", drop=False)  
print(df1_drop)

# index索引对象可以转换成列表类型
print(list(df1.index))  # ['ll', 'll', 'mm']

           x2
x1 x3        
a  11.2  11.2
b  23.2  23.2
c  23.4  23.4
a  12       1
   x1    x2    x3
x1               
a   a  11.2  11.2
b   b  23.2  23.2
c   c  23.4  23.4
a   a     1    12
[0, 1, 2, 3]

2、初始设置索引

my_df = pd.DataFrame({
    'Person': ['Alice', 'Steven', 'Neesham', 'Chris', 'Alice'],
    'City': ['Berlin', 'Montreal', 'Toronto', 'Rome', 'Munich'],
    'Mother Tongue': ['German', 'French', 'English', 'Italian', 'German'],
    'Age':  [37, 20, 38, 23, 35],

}, index=["A", "B", "C", "D", "E"])

print(my_df)

    Person      City Mother Tongue  Age
A    Alice    Berlin        German   37
B   Steven  Montreal        French   20
C  Neesham   Toronto       English   38
D    Chris      Rome       Italian   23
E    Alice    Munich        German   35

3、使用 reset_index() 方法删除 Pandas DataFrame 的索引

参数inplace:

inplace=True：不创建新的对象，直接对原始对象进行修改；
inplace=False：对数据进行修改，创建并返回新的对象承载其修改结果。

参数drop:

drop=True: 把原来的索引index列去掉，丢掉。
drop=False:保留原来的索引（以前的可能是乱的）

my_df = pd.DataFrame({
    'Person': ['Alice', 'Steven', 'Neesham', 'Chris', 'Alice'],
    'City': ['Berlin', 'Montreal', 'Toronto', 'Rome', 'Munich'],
    'Mother Tongue': ['German', 'French', 'English', 'Italian', 'German'],
    'Age':  [37, 20, 38, 23, 35],

}, index=["A", "B", "C", "D", "E"])

df_reset = my_df.reset_index()
# 等同上面
# my_df.reset_index(inplace=True)

print("Before reseting Index:")
print(my_df, "\n")

print("After reseting Index:")
print(df_reset)

Before reseting Index:
    Person      City Mother Tongue  Age
A    Alice    Berlin        German   37
B   Steven  Montreal        French   20
C  Neesham   Toronto       English   38
D    Chris      Rome       Italian   23
E    Alice    Munich        German   35 

After reseting Index:
  index   Person      City Mother Tongue  Age
0     A    Alice    Berlin        German   37
1     B   Steven  Montreal        French   20
2     C  Neesham   Toronto       English   38
3     D    Chris      Rome       Italian   23
4     E    Alice    Munich        German   35

它将重置 DataFrame 的索引，但现在的索引将显示为 index 列。如果我们想删除 index 列，我们可以在 reset_index() 方法中设置 drop=True

my_df = pd.DataFrame({
    'Person': ['Alice', 'Steven', 'Neesham', 'Chris', 'Alice'],
    'City': ['Berlin', 'Montreal', 'Toronto', 'Rome', 'Munich'],
    'Mother Tongue': ['German', 'French', 'English', 'Italian', 'German'],
    'Age':  [37, 20, 38, 23, 35],

},index=["A","B","C","D","E"])

# 重置索引，并且删除原来的索引
df_reset=my_df.reset_index(drop=True)

print("Before reseting Index:")
print(my_df,"\n")

print("After reseting Index:")
print(df_reset)

Before reseting Index:
    Person      City Mother Tongue  Age
A    Alice    Berlin        German   37
B   Steven  Montreal        French   20
C  Neesham   Toronto       English   38
D    Chris      Rome       Italian   23
E    Alice    Munich        German   35 

After reseting Index:
    Person      City Mother Tongue  Age
0    Alice    Berlin        German   37
1   Steven  Montreal        French   20
2  Neesham   Toronto       English   38
3    Chris      Rome       Italian   23
4    Alice    Munich        German   35

二、合并数据

# 以左边列为基准
pd.merge(df1, df2, how='left', on='x1')

# 以右边列为基准
pd.merge(df1, df2, how='right', on='x1')

# 已交集
pd.merge(df1, df2, how='inner', on='x1')

# 全部
pd.merge(df1, df2, how='outer', on='x1')

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	x1	x2_x	x2_y
0	a	11.2	20.23
1	b	23.2	NaN
2	c	23.4	NaN
3	d	NaN	90.4

concat拼接数据(类似append)

print(df1)

print(df2)

print('\n 列拼接\n')
print(pd.concat([df1, df2], axis=1))

print('\n 行拼接\n')
pd.concat([df1, df2])

  x111222 x211222 x311222
0       a    11.2    11.2
1       b    23.2    23.2
2       c    23.4    23.4
3       a       1      12
  x1     x2
0  a  20.23
1  b    NaN
2  d   90.4

 列拼接

  x111222 x211222 x311222   x1     x2
0       a    11.2    11.2    a  20.23
1       b    23.2    23.2    b    NaN
2       c    23.4    23.4    d   90.4
3       a       1      12  NaN    NaN

 行拼接

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	x111222	x211222	x311222	x1	x2
0	a	11.2	11.2	NaN	NaN
1	b	23.2	23.2	NaN	NaN
2	c	23.4	23.4	NaN	NaN
3	a	1	12	NaN	NaN
0	NaN	NaN	NaN	a	20.23
1	NaN	NaN	NaN	b	NaN
2	NaN	NaN	NaN	d	90.4

三、行、列的转置

原先的索引变成了行，所以说要转置时，先设置一下索引，不然转换后colums 变成了数字索引

print(df1)
# 
print(df1.T)

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
       0     1     2
x1     a     b     c
x2  11.2  23.2  23.4
x3  11.2  23.2  23.4

正确的方式

print(df1.set_index('x1'))

print('\n 转换后:\n')

print(df1.set_index('x1').T)

      x2    x3
x1            
a   11.2  11.2
b   23.2  23.2
c   23.4  23.4

 转换后:

x1     a     b     c
x2  11.2  23.2  23.4
x3  11.2  23.2  23.4

四、行列获取

1. 列获取

方式一：按标签（最常用）

print(df1)

print('获取单列')
print(df1['x1'])

print('\n获取多列\n')
print(df1[['x1','x3']])

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
3  a     1    12
获取单列
0    a
1    b
2    c
3    a
Name: x1, dtype: object

获取多列

  x1    x3
0  a  11.2
1  b  23.2
2  c  23.4
3  a    12

方式二：按索引

# 单索引
print(df1.iloc[:, 0])

# 多索引
print(df1.iloc[:, 0:])

0    a
1    b
2    c
Name: x1, dtype: object
  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4

2. 行获取

方式一：按标签获取：

df3 = df1.set_index('x1').T
print(df3)
# 按标签获取
df3.loc['x2']

x1     a     b     c
x2  11.2  23.2  23.4
x3  11.2  23.2  23.4





x1
a    11.2
b    23.2
c    23.4
Name: x2, dtype: object

方式二：按索引获取

print('正常获取：')
print(df3.iloc[0])

print('\n区间获取\n')
print(df3.iloc[0:1])

正常获取：
x1
a    11.2
b    23.2
c    23.4
Name: x2, dtype: object

区间获取

x1     a     b     c
x2  11.2  23.2  23.4

3. 精准取值

print(df1)

print('方法一：按索引精准取值')
name = df1.iloc[0, 0]  # 获取第一行，指定列数据
print("\n第一行第一列的值：" + name)

print('\n方法二：按标签精准取值\n')
name = df1.loc[0, 'x1']
print(name)
print("\n第一行第一列的值：" + name)

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
方法一：按索引精准取值

第一行第一列的值：a

方法二：按标签精准取值

a

五、行列删除

axis=0 : 行删除(默认)

axis=1 : 列删除

# print(df1)
print(df1)
print('\n 删除行\n')
print(df1.drop(['x1','x3']))

print('\n 删除列数据\n')
print(df1.drop([0], axis=1))

       0     1     2
x1     a     b     c
x2  11.2  23.2  23.4
x3  11.2  23.2  23.4

 删除行

       0     1     2
x2  11.2  23.2  23.4
       1     2
x1     b     c
x2  23.2  23.4
x3  23.2  23.4

六、列的数据格式化操作

1. 直接列操作

print(df1)

df1['x4'] = df1['x2'] + df1['x3']

print(df1)

  x1    x2    x3        x4
0  a  11.2  11.2  11.211.2
1  b  23.2  23.2  23.223.2
2  c  23.4  23.4  23.423.4
  x1    x2    x3        x4
0  a  11.2  11.2  11.211.2
1  b  23.2  23.2  23.223.2
2  c  23.4  23.4  23.423.4

2. map方式

# map 方式一
print(df1)

def test(x):
  return f'{x}1'


def test2(x):
  return f'{x}2'

df1.columns = df1.columns.map(test)
print(df1)

# 匿名函数
df1.columns = df1.columns.map(lambda x: f'{x}222')
print(df1)

  x11   x21   x31
0   a  11.2  11.2
1   b  23.2  23.2
2   c  23.4  23.4
3   a     1    12
  x111  x211  x311
0    a  11.2  11.2
1    b  23.2  23.2
2    c  23.4  23.4
3    a     1    12
  x111222 x211222 x311222
0       a    11.2    11.2
1       b    23.2    23.2
2       c    23.4    23.4
3       a       1      12

# map 方式二
df1['x11'] = df1['x11'].map({
  'a': 'a1',
  'b': 'b1',
  'c': 'c1',
})

print(df1)

   x11   x21   x31
0  NaN  11.2  11.2
1  NaN  23.2  23.2
2  NaN  23.4  23.4

3. split 字符串拆分处理

split`(*self*, *pat=None*, *n=-1*, *expand=False*)
pat：分列的依据，可以是空格，符号，字符串等等。
n：分列的次数，不指定的话就会根据符号的个数全部分列。
expand：为True可以直接将分列后的结果转换成DataFrame。
如果想要从最右边开始分列，可以使用rsplit()，rsplit()和split()的用法类似，一个从右边开始，一个从左边开始。

print(df3)
df4 = df3['brand_model'].str.split(' ', 1, expand=True)


print('\n格式化后：\n')
print(df4)

print('\n 列拼接 \n')
print(pd.concat([df3, df4], axis=1))

              brand_model cate
0            德力西 GW912200   低压
1  埃帝尔 IB-207TPN-06231C02   管阀
2         德力西 72LT3YFM40D   低压
3       科顺 2-3654-561 铆钉型   搬运

格式化后：

     0                   1
0  德力西            GW912200
1  埃帝尔  IB-207TPN-06231C02
2  德力西         72LT3YFM40D
3   科顺      2-3654-561 铆钉型

 列拼接 

              brand_model cate    0                   1
0            德力西 GW912200   低压  德力西            GW912200
1  埃帝尔 IB-207TPN-06231C02   管阀  埃帝尔  IB-207TPN-06231C02
2         德力西 72LT3YFM40D   低压  德力西         72LT3YFM40D
3       科顺 2-3654-561 铆钉型   搬运   科顺      2-3654-561 铆钉型

4. apply、applymap批量更改

print(df0)
print('\n 格式化后\n')
print(df0.apply(lambda x: x+1))

print('\n applymap格式化后\n')

print(df0.applymap(lambda x: x+2))
# 过滤将包含-改为0
df3_0 = df3.applymap(lambda x: 0 if '-' in x else x)
print(df3_0)
# 沿着列轴求和
df0[["x2"]].apply(sum, axis=0)

     x2    x3
0  11.2  11.2
1  23.2  23.2
2  23.4  23.4
3   1.0  12.0

 格式化后

     x2    x3
0  12.2  12.2
1  24.2  24.2
2  24.4  24.4
3   2.0  13.0

 applymap格式化后

     x2    x3
0  13.2  13.2
1  25.2  25.2
2  25.4  25.4
3   3.0  14.0
       brand_model cate
0     德力西 GW912200   低压
1                0   管阀
2  德力西 72LT3YFM40D   低压
3                0   搬运





x2    58.8
dtype: float64

5、类型转换

股票代码的补位，例如000123，转成csv或者excel 时，变变成123

df['正股代码']  = df['正股代码'].str.zfill(6) #用的时候必须加上.str前缀

<!-- 有时候似乎有问题 -->
df['正股代码'] = df['正股代码'].astype('str') #将原本的int数据类型转换为文本

df.to_csv('./test.csv', dtype = object)

6、变更类型

dfx4 = pd.DataFrame([['11', '1.2', '3'], ['22', '4.8', '5']], columns=['a', 'b', 'c'])

# dfx4['a'].str.replace('万','').astype('int')
dfx4['b'].astype('int')
dfx4['b'].dtypes
# dfx4['b']

---------------------------------------------------------------------------

ValueError                                Traceback (most recent call last)

/var/folders/w6/9k4dzqlj617f06dfby_vk1pr0000gn/T/ipykernel_14591/716864053.py in <module>
      2 
      3 # dfx4['a'].str.replace('万','').astype('int')
----> 4 dfx4['b'].astype('int')
      5 dfx4['b'].dtypes
      6 # dfx4['b']


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/generic.py in astype(self, dtype, copy, errors)
   5813         else:
   5814             # else, only a single dtype is given
-> 5815             new_data = self._mgr.astype(dtype=dtype, copy=copy, errors=errors)
   5816             return self._constructor(new_data).__finalize__(self, method="astype")
   5817 


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/internals/managers.py in astype(self, dtype, copy, errors)
    416 
    417     def astype(self: T, dtype, copy: bool = False, errors: str = "raise") -> T:
--> 418         return self.apply("astype", dtype=dtype, copy=copy, errors=errors)
    419 
    420     def convert(


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/internals/managers.py in apply(self, f, align_keys, ignore_failures, **kwargs)
    325                     applied = b.apply(f, **kwargs)
    326                 else:
--> 327                     applied = getattr(b, f)(**kwargs)
    328             except (TypeError, NotImplementedError):
    329                 if not ignore_failures:


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/internals/blocks.py in astype(self, dtype, copy, errors)
    589         values = self.values
    590 
--> 591         new_values = astype_array_safe(values, dtype, copy=copy, errors=errors)
    592 
    593         new_values = maybe_coerce_values(new_values)


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/dtypes/cast.py in astype_array_safe(values, dtype, copy, errors)
   1307 
   1308     try:
-> 1309         new_values = astype_array(values, dtype, copy=copy)
   1310     except (ValueError, TypeError):
   1311         # e.g. astype_nansafe can fail on object-dtype of strings


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/dtypes/cast.py in astype_array(values, dtype, copy)
   1255 
   1256     else:
-> 1257         values = astype_nansafe(values, dtype, copy=copy)
   1258 
   1259     # in pandas we don't store numpy str dtypes, so convert to object


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/core/dtypes/cast.py in astype_nansafe(arr, dtype, copy, skipna)
   1172         # work around NumPy brokenness, #1987
   1173         if np.issubdtype(dtype.type, np.integer):
-> 1174             return lib.astype_intsafe(arr, dtype)
   1175 
   1176         # if we have a datetime/timedelta array of objects


/usr/local/anaconda3/envs/py39/lib/python3.9/site-packages/pandas/_libs/lib.pyx in pandas._libs.lib.astype_intsafe()


ValueError: invalid literal for int() with base 10: '1.2'

dfx3 = pd.DataFrame([['11', '1.2', '3'], ['22', '4.8', '5']], columns=['a', 'b', 'c'])
print(dfx3)
# pd.to_numeric()自动转换成float类型
print(dfx3.apply(pd.to_numeric, errors='ignore', downcast='float'))
print(dfx3.dtypes)
# 遍历的方式格式转换
dfx3.applymap(lambda x: float(x))
# print(dfx3['a'].dtypes)


# df['a'].str.replace('万','').astype('float')

    a    b  c
0  11  1.2  3
1  22  4.8  5
      a    b    c
0  11.0  1.2  3.0
1  22.0  4.8  5.0
a    object
b    object
c    object
dtype: object

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	a	b	c
0	11.0	1.2	3.0
1	22.0	4.8	5.0

7、查看列类型

print(dfx3['a'].dtypes)

七、行、列名称更改

1. 单独更改

不会更改源数据，要重新赋值

print(df1)

print('\n---指定行更改--- \n')

print(df1.rename({ 'x1': 'x11' }))

print('\n---指定行、列更改--- \n')

df1.rename(index={'x2': 'bj'}, columns={
                 0: 11})  # 为某个 index 单独修改名称

       0     1     2
x1     a     b     c
x2  11.2  23.2  23.4
x3  11.2  23.2  23.4

---指定行更改--- 

        0     1     2
x11     a     b     c
x2   11.2  23.2  23.4
x3   11.2  23.2  23.4

---指定行、列更改---

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	11	1	2
x1	a	b	c
bj	11.2	23.2	23.4
x3	11.2	23.2	23.4

2. 批量更改

print(df1)

# 自定义map函数
def test_map(x):
    return x+'_ABC'


# print(df1.index.map(test_map))
# 输出 Index(['BEIJING_ABC', 'SHANGHAI_ABC', 'GUANGZHOU_ABC'], dtype='object')

# print(df1.rename(index=test_map))
print(df1.rename(index=lambda x: f'{x}%'))

# print(df1['x1'].map(lambda x: f'{x}%'))
# df1['x1'] = df1['x1'].map(lambda x: f'{x}%')
# print(df1)

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
3  a     1    12
   x1    x2    x3
0%  a  11.2  11.2
1%  b  23.2  23.2
2%  c  23.4  23.4
3%  a     1    12
   x1    x2    x3
0  a%  11.2  11.2
1  b%  23.2  23.2
2  c%  23.4  23.4
3  a%     1    12

批量去除%,并转化为float类型（排序）

print(df6)
df6_out = df6.applymap(lambda x: x if not x.endswith('%') else float(x.replace('%', '')))
print(df6_out)

    x1     x2     x3
0  12%   11.2   11.2
1   23  23.2%  23.2%
2   99   23.4  23.4%
3  123      1    12%
     x1    x2    x3
0  12.0  11.2  11.2
1    23  23.2  23.2
2    99  23.4  23.4
3   123     1  12.0

八、去重

data.drop_duplicates(subset=['A','B'],keep='first',inplace=True)

subset对应的值是列名，表示只考虑这两列，将这两列对应值相同的行进行去重。默认值为subset=None表示考虑所有列。
keep='first'表示保留第一次出现的重复行，是默认值。keep另外两个取值为"last"和False，分别表示保留最后一次出现的重复行和去除所有重复行。
inplace=True表示直接在原来的DataFrame上删除重复项，而默认值False表示生成一个副本

print(df1)
print('去重')
df1 = df1.drop_duplicates(['x1'])
print(df1)

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
去重
  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4

九、累加

DataFrame.cumsum(axis=None, skipna=True, *args, **kwargs)

axis 对每行的值累加（1或columns）或者对每列的值累加（0或index）可选值{0 or ‘index’, 1 or ‘columns’}, 默认值0
skipna 是否排除空值（NaN）默认值True
返回值为Series 或 DataFrame

print(df1)

print('行累加')

# 行累加
print(df1.cumsum())

print('列累加')

print(df1.cumsum(axis=1))

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
3  a     1    12
行累加
     x1             x2              x3
0     a           11.2            11.2
1    ab       11.223.2        11.223.2
2   abc   11.223.223.4    11.223.223.4
3  abca  11.223.223.41  11.223.223.412
列累加
  x1     x2         x3
0  a  a11.2  a11.211.2
1  b  b23.2  b23.223.2
2  c  c23.4  c23.423.4
3  a     a1       a112

十、排序

print(df5)
# 先转化成number
df5['x1'] = df5['x1'].map(lambda x: float(x))
df5.sort_values(by='x1') # 默认升序

      x1    x2    x3
0   12.0  11.2  11.2
1   23.0  23.2  23.2
2   99.0  23.4  23.4
3  123.0     1    12

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	x1	x2	x3
0	12.0	11.2	11.2
1	23.0	23.2	23.2
2	99.0	23.4	23.4
3	123.0	1	12

删除指定条件的行

print(df5)
# 先转成数字类型
df5= df5.applymap(lambda x: float(x))
# 删除x1列大于20的第一行
print(df5.drop(index=df5[df5['x1'] > 20].index[0]))
# 删除x1列大于20的所有行
print(df5.drop(index=df5[df5['x1'] > 20].index))

      x1    x2    x3
0   12.0  11.2  11.2
1   23.0  23.2  23.2
2   99.0  23.4  23.4
3  123.0   1.0  12.0
      x1    x2    x3
0   12.0  11.2  11.2
2   99.0  23.4  23.4
3  123.0   1.0  12.0
     x1    x2    x3
0  12.0  11.2  11.2

dataframe to json

print(df5)

df5.to_json(orient='table')

    x1    x2    x3
0   12  11.2  11.2
1   23  23.2  23.2
2   99  23.4  23.4
3  123     1    12





'{"schema":{"fields":[{"name":"index","type":"integer"},{"name":"x1","type":"string"},{"name":"x2","type":"string"},{"name":"x3","type":"string"}],"primaryKey":["index"],"pandas_version":"0.20.0"},"data":[{"index":0,"x1":"12","x2":"11.2","x3":"11.2"},{"index":1,"x1":"23","x2":"23.2","x3":"23.2"},{"index":2,"x1":"99","x2":"23.4","x3":"23.4"},{"index":3,"x1":"123","x2":"1","x3":"12"}]}'

格式化

转换成数字类型

print(df1)
# print(df1['x2'])
df1['x2'] = df1['x2'].apply(pd.to_numeric)
# df = df.apply(pd.to_numeric, errors='ignore', downcast='float')
print('格式化')
print(df1['x2'])

  x1    x2    x3
0  a  11.2  11.2
1  b  23.2  23.2
2  c  23.4  23.4
3  a   1.0    12
格式化
0    11.2
1    23.2
2    23.4
3     1.0
Name: x2, dtype: float64

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

pandas.dataframe常用操作.md

pandas.dataframe常用操作.md

dataframe 的常见操作

一、索引的最常见操作

1、设置索引

2、初始设置索引

3、使用 reset_index() 方法删除 Pandas DataFrame 的索引

二、合并数据

concat拼接数据(类似append)

三、行、列的转置

正确的方式

四、行列获取

1. 列获取

2. 行获取

3. 精准取值

五、行列删除

六、列的数据格式化操作

1. 直接列操作

2. map方式

3. split 字符串拆分处理

4. apply、applymap批量更改

5、类型转换

6、变更类型

7、查看列类型

七、行、列名称更改

1. 单独更改

2. 批量更改

批量去除%,并转化为float类型（排序）

八、去重

九、累加

十、排序

删除指定条件的行

dataframe to json

格式化

转换成数字类型

Files

pandas.dataframe常用操作.md

Latest commit

History

pandas.dataframe常用操作.md

File metadata and controls

dataframe 的常见操作

一、索引的最常见操作

1、 设置索引

2、 初始设置索引

3、 使用 reset_index() 方法删除 Pandas DataFrame 的索引

二、合并数据

concat拼接数据(类似append)

三、行、列的转置

正确的方式

四、行列获取

1. 列获取

2. 行获取

3. 精准取值

五、行列删除

六、列的数据格式化操作

1. 直接列操作

2. map方式

3. split 字符串拆分处理

4. apply、applymap批量更改

5、类型转换

6、变更类型

7、查看列类型

七、行、列名称更改

1. 单独更改

2. 批量更改

批量去除%,并转化为float类型（排序）

八、去重

九、累加

十、排序

删除指定条件的行

dataframe to json

格式化

转换成数字类型

1、设置索引

2、初始设置索引

3、使用 reset_index() 方法删除 Pandas DataFrame 的索引