根据特定列合并多个 CSV 文件 - Python

Question

我正在尝试在 Pandas 中合并大约 101 个 CSV 文件。每个文件都有 2 个时间列和一个“值”列。我想保留 2 次列，因为它们在 CSV 文件中是相同的，然后将 101 个 CSV 中的每一个的“值”列合并到一个新的 CSV 文件中。

使用 pd.merge 我可以使用以下合并 2 个文件

data1 = {'time': ['00:00','01:00','02:00'], 
        'local_time': ['09:30','10:30','11:30'],
        'value': ['265.591','330.766','360.962']}

data2 = {'time': ['00:00','01:00','02:00'], 
        'local_time': ['09:30','10:30','11:30'],
        'value': ['521.217','588.034','588.034']}

df_1 = pd.DataFrame(data1)
df_2 = pd.DataFrame(data2)
locs = ['_A11','_B10']

df_test = pd.merge(df_1,df_2, on=['time','local_time'], how='inner', suffixes = (locs)
)

print(df_test)

这会产生：

    time local_time value_A11 value_B10
0  00:00      09:30   265.591   521.217
1  01:00      10:30   330.766   588.034
2  02:00      11:30   360.962   588.034

但是，我不太确定如何组合接下来的 99 个 csv 文件，或者这是否是完成这项任务的最佳方式。

我的目标是：

    time local_time value_A11 value_B10 value_B11 ...
0  00:00      09:30   265.591   521.217       123 ...
1  01:00      10:30   330.766   588.034       456 ...
2  02:00      11:30   360.962   588.034       789 ...

任何帮助将不胜感激！

编辑 1：

Colin 的示例有效，但是我一直在将数据帧加载到这样的数组中：

import glob
import os

# create and sort list of file names in folder
fl = glob.glob('*.csv')
sorted_fl = sorted(fl)

# open csv files from list and store in df
df_list = [pd.read_csv(f, header=3) for f in sorted_fl]

#test df
df_list[0]

我想知道如何修改 for 循环以便它可以为数组提供数据？再次感谢！

编辑 2：从答案到编辑 1 的错误

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-144-772c1d15f228> in <module>
     14 # loop through each dataframe and merge it with existing one
     15 for i, df in enumerate(df_list[1:]):
---> 16   df_output = pd.merge(df_list[0], df, on=['time','local_time'], how='inner', suffixes = (['_' + str(i), '_' + str(i+1)]))

~/opt/anaconda3/lib/python3.7/site-packages/pandas/core/reshape/merge.py in merge(left, right, how, on, left_on, right_on, left_index, right_index, sort, suffixes, copy, indicator, validate)
     79         copy=copy,
     80         indicator=indicator,
---> 81         validate=validate,
     82     )
     83     return op.get_result()

~/opt/anaconda3/lib/python3.7/site-packages/pandas/core/reshape/merge.py in __init__(self, left, right, how, on, left_on, right_on, axis, left_index, right_index, sort, suffixes, copy, indicator, validate)
    628         # validate the merge keys dtypes. We may need to coerce
    629         # to avoid incompat dtypes
--> 630         self._maybe_coerce_merge_keys()
    631 
    632         # If argument passed to validate,

~/opt/anaconda3/lib/python3.7/site-packages/pandas/core/reshape/merge.py in _maybe_coerce_merge_keys(self)
   1136                     inferred_right in string_types and inferred_left not in string_types
   1137                 ):
-> 1138                     raise ValueError(msg)
   1139 
   1140             # datetimelikes must match exactly

ValueError: You are trying to merge on object and float64 columns. If you wish to proceed you should use pd.concat

编辑 3

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-3-cce982321079> in <module>
     11 
     12 # change datatype to datetime for first df
---> 13 df['local_time'] = pd.to_datetime(df_list[0]['local_time'])
     14 df['time'] = pd.to_datetime(df_list[0]['time'])
     15 

NameError: name 'df' is not defined

Answer 1

这似乎是一个不错的方法。我只是设置了一些不同的合并和后缀，这样你就可以遍历每个数据帧，如下所示。每个新值列都将合并到 df_test。

编辑：更新代码以配合 OP 的编辑

编辑 2： 修复了 OP 错误的数据类型

import pandas as pd    
import glob
import os

# create and sort list of file names in folder
fl = glob.glob('*.csv')
sorted_fl = sorted(fl)

# open csv files from list and store in df
df_list = [pd.read_csv(f, header=3) for f in sorted_fl]

# change datatype to datetime for first df
df['local_time'] = pd.to_datetime(df_list[0]['local_time'])
df['time'] = pd.to_datetime(df_list[0]['time'])


# loop through each dataframe and merge it with existing one
for i, df in enumerate(df_list[1:]):

  # change datatype to datetime
  df['local_time'] = pd.to_datetime(df['local_time'])
  df['time'] = pd.to_datetime(df['time'])

  df_output = pd.merge(df_list[0], df, on=['time','local_time'], how='inner', suffixes = (['_' + str(i), '_' + str(i+1)]))

#print(df_output)
'''
    time local_time  value_0  value_1  value_2  value_3
0  00:00      09:30  738.591  265.591  521.217  856.217
1  01:00      10:30  217.766  330.766  588.034  346.034
2  02:00      11:30  295.962  360.962  588.034  645.034
'''