如何以更好的方式聚合多个 DataFrame？答案

【问题标题】：How To Aggregate Multiple DataFrame in a Better Way?如何以更好的方式聚合多个 DataFrame？
【发布时间】：2019-03-13 21:54:13
【问题描述】：

所以我有 36 个 DataFrame 我必须在这三列 ['Sensor ID' , 'Time Instant' , 'Measurement'] 上合并到一个 DataFrame 中]

所以，这就是我所做的：

mi_pollution_1 = pd.read_csv('/content/drive/My Drive/DatiAirQuality/MI_Air_Quality/data/mi_pollution_10273.csv' )
mi_pollution_1.columns= ['Sensor ID' , 'Time Instant' , 'Measurement']
mi_pollution_2 = pd.read_csv('/content/drive/My Drive/DatiAirQuality/MI_Air_Quality/data/mi_pollution_10278.csv')
mi_pollution_2.columns= ['Sensor ID' , 'Time Instant' , 'Measurement']
mi_pollution_3 = pd.read_csv('/content/drive/My Drive/DatiAirQuality/MI_Air_Quality/data/mi_pollution_10279.csv')
mi_pollution_3.columns= ['Sensor ID' , 'Time Instant' , 'Measurement']

.
.
.
.
mi_pollution_35= pd.read_csv('/content/drive/My Drive/DatiAirQuality/MI_Air_Quality/data/mi_pollution_6372.csv')
mi_pollution_35.columns= ['Sensor ID' , 'Time Instant' , 'Measurement']
mi_pollution_36= pd.read_csv('/content/drive/My Drive/DatiAirQuality/MI_Air_Quality/data/mi_pollution_6372.csv')
mi_pollution_36.columns= ['Sensor ID' , 'Time Instant' , 'Measurement']

然后：

 frames = [mi_pollution_1 , mi_pollution_2 ,mi_pollution_3 , mi_pollution_4,
     mi_pollution_5 , mi_pollution_6,mi_pollution_7 , mi_pollution_8,
     mi_pollution_9 , mi_pollution_10,mi_pollution_11 , mi_pollution_12, 
     mi_pollution_13 , mi_pollution_14, mi_pollution_15 , mi_pollution_16,
     mi_pollution_17 , mi_pollution_18 ,mi_pollution_19 , mi_pollution_20,
     mi_pollution_21 , mi_pollution_22,mi_pollution_23 , mi_pollution_24,
     mi_pollution_25 , mi_pollution_26,mi_pollution_27 , mi_pollution_28, 
     mi_pollution_29 , mi_pollution_30, mi_pollution_31 , mi_pollution_32,
     mi_pollution_33 , mi_pollution_34, mi_pollution_35 , mi_pollution_36]

 df_result = pd.merge(frames , on = ['Sensor ID' , 'Time Instant' , 'Measurement'])

所以，我想知道是否有更有效和“更清洁”的方式来实现它。谢谢

【问题讨论】：

标签： python-3.x dataframe merge

【解决方案1】：

如果您的文件内容是同质的，看起来它们是基于您上面的示例，我认为首先将数据加载为列表可能会更干净。

import csv
import pandas as pd

file_name = ['sample_file1.csv', 'sample_file2.csv', 'sample_file3.csv']

content = []
for n in file_name:
    with open(n) as f: 
        reader = csv.reader(f, delimiter=',')
        content += list(reader)[1:] # If the first row contains headers

df = pd.DataFrame(content)
df.columns = ['Sensor ID' , 'Time Instant' , 'Measurement']

【讨论】：

非常感谢，确实，比上面的“混乱”更干净有效:-)，再次感谢