如何使用 Python Pandas 创建具有相同数据库的多个数据框

Question

第 1 步：我从 csv 文件中导入了带有 pandas 的数据库：

import pandas as pd
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
path_file = os.path.join(BASE_DIR, 'csv/complete_kiko_camel_product_list.csv')
df_read = pd.read_csv(path_file)

第 2 步：我想使用所有数据创建多个数据框，但具有不同的特征：

按 column_name_1 排序的前 1000 个元素

按 column_name_2 排序的前 1000 个元素

和 go on（总共 5 个数据帧）

我尝试过这样的事情，但它会引发错误，因为数据框不表现为变量：

df_text = df_read.sort_values('ratio_text', axis = 0, ascending = False, inplace = True, na_position ='last').head(1000)
df_ssim = df_read.sort_values('ratio_image', axis = 0, ascending = False, inplace = True, na_position ='last').head(1000)
df_hist = df_read.sort_values('average_histogram_ssim', axis = 0, ascending = False, inplace = True, na_position ='last').head(1000)
df_sust = df_read.sort_values('average_substract_ssim', axis = 0, ascending = False, inplace = True, na_position ='last').head(1000) 

第 3 步：最后我想创建一个连接其他 5 个的数据框：

frames = [df_gray, df_text, df_ssim, df_hist, df_sust]
df_ordered = pd.concat(frames)

如何修复第 2 步？

Answer 1

您应该摆脱所有inplace=True并且因为axes=0是默认值，您可以将其删除。

Answer 2

创建dataframes的dict并concat dict

综合数据

import pandas as pd
import numpy as np

np.random.seed(365)
data = {'gray': [np.random.randint(10) for _ in range(2000)],
        'ratio_text': [np.random.randint(10) for _ in range(2000)],
        'ratio_image': [np.random.randint(10) for _ in range(2000)],
        'avg_hist_ssim': [np.random.randint(10) for _ in range(2000)],
        'avg_sub_ssim': [np.random.randint(10) for _ in range(2000)]}


df = pd.DataFrame(data)
df.head()

 gray  ratio_text  ratio_image  avg_hist_ssim  avg_sub_ssim
    2           4            3              1             7
    4           2            3              1             7
    1           5            2              3             2
    5           1            7              5             5
    2           4            8              5             5

有序数据帧的dict

df_dict = {col: df.sort_values(col, ascending=False, na_position ='last').head(1000) for col in df.columns}

for k, v in df_dict.items():
    print(f'DataFrame: {k}\n{v.head()}\n')

DataFrame: gray
      gray  ratio_text  ratio_image  avg_hist_ssim  avg_sub_ssim
567      9           4            9              7             4
709      9           5            6              1             9
520      9           3            3              1             6
713      9           4            3              3             7
1636     9           3            8              4             0

DataFrame: ratio_text
      gray  ratio_text  ratio_image  avg_hist_ssim  avg_sub_ssim
1460     9           9            5              3             5
1253     6           9            7              3             2
350      8           9            6              2             2
964      4           9            9              8             3
1027     4           9            9              1             1

DataFrame: ratio_image
      gray  ratio_text  ratio_image  avg_hist_ssim  avg_sub_ssim
1031     4           3            9              6             5
144      9           0            9              6             5
271      6           3            9              1             5
1847     8           0            9              7             1
1053     3           3            9              3             0

DataFrame: avg_hist_ssim
      gray  ratio_text  ratio_image  avg_hist_ssim  avg_sub_ssim
1458     7           3            4              9             6
829      2           6            6              9             9
796      2           4            4              9             0
1929     0           7            2              9             0
808      5           1            6              9             8

DataFrame: avg_sub_ssim
      gray  ratio_text  ratio_image  avg_hist_ssim  avg_sub_ssim
280      0           6            0              0             9
847      5           1            0              4             9
1314     7           1            6              9             9
821      1           1            9              2             9
1305     6           9            7              8             9

df_dict concat

df_ordered = pd.concat(df_dict)