读取由 nan 行拆分的数据帧，并在 Python 中将它们重塑为多个数据帧

Question

我有一个来自here 的示例excel 文件data1.xlsx，它有一个Sheet1，如下所示：

现在我想用openpyxl 或pandas 阅读，然后将它们转换为新的df1 和df2，最后我将它们保存为price 和quantity 表：

价格表：

和数量表

我使用过的代码：

df = pd.read_excel('./data1.xlsx', sheet_name = 'Sheet1')
df_list = np.split(df, df[df.isnull().all(1)].index) 

for df in df_list:
    print(df, '\n')

输出：

         bj  Unnamed: 1  Unnamed: 2  Unnamed: 3 Unnamed: 4
0      year      2018.0      2019.0      2020.0        sum
1     price        12.0         4.0         5.0         21
2  quantity         5.0         5.0         3.0         13 

         bj  Unnamed: 1  Unnamed: 2  Unnamed: 3 Unnamed: 4
3       NaN         NaN         NaN         NaN        NaN
4        sh         NaN         NaN         NaN        NaN
5      year      2018.0      2019.0      2020.0        sum
6     price         5.0         6.0         7.0         18
7  quantity         7.0         5.0         4.0         16 

    bj  Unnamed: 1  Unnamed: 2  Unnamed: 3 Unnamed: 4
8  NaN         NaN         NaN         NaN        NaN 

          bj  Unnamed: 1  Unnamed: 2  Unnamed: 3 Unnamed: 4
9        NaN         NaN         NaN         NaN        NaN
10        gz         NaN         NaN         NaN        NaN
11      year      2018.0      2019.0      2020.0        sum
12     price         2.0         3.0         1.0          6
13  quantity         6.0         9.0         3.0         18 

     bj  Unnamed: 1  Unnamed: 2  Unnamed: 3 Unnamed: 4
14  NaN         NaN         NaN         NaN        NaN 

          bj  Unnamed: 1  Unnamed: 2  Unnamed: 3 Unnamed: 4
15       NaN         NaN         NaN         NaN        NaN
16        sz         NaN         NaN         NaN        NaN
17      year      2018.0      2019.0      2020.0        sum
18     price         8.0         2.0         3.0         13
19  quantity         5.0         4.0         3.0         12 

我如何在 Python 中做到这一点？非常感谢。

Answer 1

用途：

#add header=None for default columns names
df = pd.read_excel('./data1.xlsx', sheet_name = 'Sheet1', header=None)

#convert columns by second row
df.columns = df.iloc[1].rename(None)

#create new column `city` by forward filling non missing values by second column
df.insert(0, 'city', df.iloc[:, 0].mask(df.iloc[:, 1].notna()).ffill())
#convert floats to integers 
df.columns = [int(x) if isinstance(x, float) else x for x in df.columns]
#convert column year to index
df = df.set_index('year')

---

print (df)
         city    2018    2019    2020  sum
year                                      
bj         bj     NaN     NaN     NaN  NaN
year       bj  2018.0  2019.0  2020.0  sum
price      bj    12.0     4.0     5.0   21
quantity   bj     5.0     5.0     3.0   13
NaN        bj     NaN     NaN     NaN  NaN
sh         sh     NaN     NaN     NaN  NaN
year       sh  2018.0  2019.0  2020.0  sum
price      sh     5.0     6.0     7.0   18
quantity   sh     7.0     5.0     4.0   16
NaN        sh     NaN     NaN     NaN  NaN
NaN        sh     NaN     NaN     NaN  NaN
gz         gz     NaN     NaN     NaN  NaN
year       gz  2018.0  2019.0  2020.0  sum
price      gz     2.0     3.0     1.0    6
quantity   gz     6.0     9.0     3.0   18
NaN        gz     NaN     NaN     NaN  NaN
NaN        gz     NaN     NaN     NaN  NaN
sz         sz     NaN     NaN     NaN  NaN
year       sz  2018.0  2019.0  2020.0  sum
price      sz     8.0     2.0     3.0   13
quantity   sz     5.0     4.0     3.0   12

---

df1 = df.loc['price'].reset_index(drop=True)
print (df1)
  city  2018  2019  2020 sum
0   bj  12.0   4.0   5.0  21
1   sh   5.0   6.0   7.0  18
2   gz   2.0   3.0   1.0   6
3   sz   8.0   2.0   3.0  13

df2 = df.loc['quantity'].reset_index(drop=True)
print (df2)
  city  2018  2019  2020 sum
0   bj   5.0   5.0   3.0  13
1   sh   7.0   5.0   4.0  16
2   gz   6.0   9.0   3.0  18
3   sz   5.0   4.0   3.0  12

最后一次将DataFrames 写入现有文件可以通过mode='a' 参数link：

with pd.ExcelWriter('data1.xlsx', mode='a') as writer:  
    df1.to_excel(writer, sheet_name='price')
    df2.to_excel(writer, sheet_name='quantity')