1

使用 pathlib.Path().glob()，我们如何遍历一个目录并在每次迭代时读入 2 个文件？

假设我的目录C:\Users\server\Desktop\Dataset如下所示：

P1_mean_fle.csv
P2_mean_fle.csv
P3_mean_fle.csv
P1_std_dev_fle.csv
P2_std_dev_fle.csv
P3_std_dev_fle.csv

如果我只想在 Pi 的每次迭代中读取 1 个文件，我的代码将如下所示：

from pathlib import Path
import pandas as pd

file_path = r'C:\Users\server\Desktop\Dataset'
param_file = 'P*' + '_mean_fle.csv'

for i, fle in enumerate(Path(file_path).glob(param_file)):
    mean_fle = pd.read_csv(fle).values

    results = tuning(mean_fle)  #tuning is some function which takes in the file mean 
                                #and does something with this file

现在，我如何在 Pi 的每次迭代中读取 2 个文件？下面的代码不太有效，因为param_file只能分配一种文件名类型。如果有办法使用pathlib.

from pathlib import Path
import pandas as pd

param_file = 'P*' + '_mean_fle.csv'
param_file = 'P*' + '_std_dev_fle.csv'  #this is wrong

for i, fle in enumerate(Path(file_path).glob(param_file)):  #this is wrong inside the glob() part
    mean_fle = pd.read_csv(fle).values
    std_dev_fle = pd.read_csv(fle).values

    results = tuning(mean_fle, std_dev_fle)  #tuning is some function which takes in the two files mean 
                                             #and std_dev and does something with these 2 files

先感谢您。

4

2 回答 2

3

如果您的文件名遵循示例中的确定性规则，则最好的办法是迭代一种文件，并通过字符串替换找到相应的文件。

from pathlib import Path
import pandas as pd

file_path = r'C:\Users\server\Desktop\Dataset'
param_file = 'P*' + '_mean_fle.csv'

for i, fle in enumerate(Path(file_path).glob(param_file)):
    stddev_fle = fle.with_name(fle.name.replace("mean", "std_dev"))
    mean_values = pd.read_csv(fle).values
    stddev_values = pd.read_csv(stddev_fle).values

    results = tuning(mean_values, stddev_values)
于 2020-04-15T04:53:40.850 回答
1

我建议你两种方法：

1.

如果您确定所有文件的编号都没有“漏洞”，则可以只使用没有“glob”的文件：

mean_csv_pattern = 'P{}_mean_fle.csv'
std_dev_pattern = 'P{}_std_dev_fle.csv'

i = 0
while True:
    i += 1
    try:
        mean_fle = pd.read_csv(mean_csv_pattern.format(i)).values
        std_dev_fle = pd.read_csv(std_dev_pattern.format(i)).values
    except (<put your exceptions here>):
        break
    results = tuning(mean_fle, std_dev_fle)

2.

使用预取操作获取所有文件并将它们放入可以在主循环中查询的结构中。

Glob 用于平均文件，glob 用于 std_dev 文件，从文件名中获取数字并生成字典 {index: {'mean_file': mean_file, 'std_file': std_file)} 然后循环排序字典键......

于 2020-04-15T04:52:59.393 回答