如何在Python中的csv的所有列中获取特定类别的结果的datetime值？

格伦·韦加斯

我有一个包含20天用户社交媒体活动值的csv文件，我想获取第1天用户活动的详细信息，这是csv中条目的示例


DateTime                  Instagram  Facebook   Twitter
(2020,09,01,10,00,00)        Y          N         Y
(2020,09,01,10,01,00)        N          Y         Y
(2020,09,01,10,02,00)        N          Y         N
(2020,09,01,10,03,00)        N          Y         N
(2020,09,01,10,04,00)        Y          N         Y
(2020,09,01,11,00,00)        Y          N         N
(2020,09,02,10,00,00)        N          Y         Y
(2020,09,02,10,00,00)        Y          N         N
(2020,09,02,10,00,00)        N          N         N
(2020,09,03,10,00,00)        Y          Y         Y

Y代表用户处于活动状态，N代表用户处于非活动状态我想显示2020-09-01第一天（2020年9月1日）所有应用程序的活动状态。

所以我希望结果看起来像这样（仅当用户在该应用上处于活动状态（Y）时的日期时间值）

{'Instagram':[(2020,09,01,10,00,00),(2020,09,01,10,04,00),(2020,09,01,11,00,00)],
 'Facebook':[(2020,09,01,10,01,00), (2020,09,01,10,02,00), (2020,09,01,10,03,00)],
 'Twitter':[(2020,09,01,10,00,00), (2020,09,01,10,01,00), (2020,09,01,10,04,00)]}

我写了一段代码，但是没有给我想要的结果

df['DateTime'] = pd.to_datetime(df['DateTime'], format='(%Y,%m,%d,%H,%M,%S)')

for idx, d in df.groupby(df['DateTime'].dt.date):
    print(d.drop('DateTime', axis=1).to_dict('list'))

This was the result I got
    
{'Instagram': ['Y', 'N', 'N', 'N', 'Y', 'Y'], 'Facebook': ['N', 'Y', 'Y', 'Y', 'N', 'N'], 'Twitter': ['Y', 'Y', 'N', 'N', 'Y', 'N']}
{'Instagram': ['N', 'Y', 'N'], 'Facebook': ['Y', 'N', 'N'], 'Twitter': ['Y', 'N', 'N']}
{'Instagram': ['Y'], 'Facebook': ['Y'], 'Twitter': ['Y']}

DateTime列包含datetime对象格式的值，我将其转换为pandas datetime格式

耶斯列尔

将值转换为新列，按过滤第一个日期boolean indexing，然后按DataFrame.melt和取消过滤list：

df['d'] = pd.to_datetime(df['DateTime'], format='(%Y,%m,%d,%H,%M,%S)')

day1 = df['d'].dt.date[0]
df = df[df['d'].dt.date.eq(day1)] 

df = df.melt(['DateTime','d']) 
df = df[df['value'].eq('Y')] 


d = df.groupby('variable')['DateTime'].agg(list).to_dict()
print (d)
{'Facebook': ['(2020,09,01,10,01,00)', '(2020,09,01,10,02,00)', '(2020,09,01,10,03,00)'], 
 'Instagram': ['(2020,09,01,10,00,00)', '(2020,09,01,10,04,00)', '(2020,09,01,11,00,00)'], 
 'Twitter': ['(2020,09,01,10,00,00)', '(2020,09,01,10,01,00)', '(2020,09,01,10,04,00)']}

如果需要每个datetime嵌套字典的输出：

df['d'] = pd.to_datetime(df['DateTime'], format='(%Y,%m,%d,%H,%M,%S)')

df = df.melt(['DateTime','d']) 
df = df[df['value'].eq('Y')] 

s = df.groupby([df['d'].dt.strftime('%Y-%m-%d'), 'variable'])['DateTime'].agg(list)
print (s)

d1 = {level: s.xs(level).to_dict() for level in s.index.levels[0]}
print (d1)
{'2020-09-01': {'Facebook': ['(2020,09,01,10,01,00)', '(2020,09,01,10,02,00)', '(2020,09,01,10,03,00)'], 
                'Instagram': ['(2020,09,01,10,00,00)', '(2020,09,01,10,04,00)', '(2020,09,01,11,00,00)'],
                'Twitter': ['(2020,09,01,10,00,00)', '(2020,09,01,10,01,00)', '(2020,09,01,10,04,00)']},
 '2020-09-02': {'Facebook': ['(2020,09,02,10,00,00)'], 
                'Instagram': ['(2020,09,02,10,00,00)'], 
                'Twitter': ['(2020,09,02,10,00,00)']}, 
 '2020-09-03': {'Facebook': ['(2020,09,03,10,00,00)'], 
                'Instagram': ['(2020,09,03,10,00,00)'], 
                'Twitter': ['(2020,09,03,10,00,00)']}}

print (d1['2020-09-01'])
{'Facebook': ['(2020,09,01,10,01,00)', '(2020,09,01,10,02,00)', '(2020,09,01,10,03,00)'], 
 'Instagram': ['(2020,09,01,10,00,00)', '(2020,09,01,10,04,00)', '(2020,09,01,11,00,00)'],
 'Twitter': ['(2020,09,01,10,00,00)', '(2020,09,01,10,01,00)', '(2020,09,01,10,04,00)']}

print (d1['2020-09-02'])
{'Facebook': ['(2020,09,02,10,00,00)'], 'Instagram': ['(2020,09,02,10,00,00)'], 'Twitter': ['(2020,09,02,10,00,00)']}

本文收集自互联网，转载请注明来源。

如有侵权，请联系 [email protected] 删除。