使用Pandas，如何按天分组JSON数据并计算出现次数？

Question

I know this question looks a bit absurd, but I've been searching for its solution all day long today and got nothing.

I have a JSON object

times = [1554278809, 1554276258, 1554274173, 1554270457] /* four occurrences in April 3 */

I need to group it by day and count the occurrence and get a result like:

result = [{'2019-04-01': 0, '2019-04-02': 0, '2019-04-03': 4}]

So far, I've tried this (which, as expected, doesn't work):

times = [str(datetime.fromtimestamp(x)) for x in times]
df = pd.DataFrame(times)
print(df.groupby(df.index.date).count())

Answer 1

Create Series with to_datetime convert unix time and then convert it to strings by Series.dt.strftime :

times = [1554278809, 1554276258, 1554274173, 1554270457] 
s = pd.Series(pd.to_datetime(times, unit='s'))
print (s)
0   2019-04-03 08:06:49
1   2019-04-03 07:24:18
2   2019-04-03 06:49:33
3   2019-04-03 05:47:37
dtype: datetime64[ns]

print(s.groupby(s.dt.strftime('%Y-%m-%d')).count().to_dict())
{'2019-04-03': 4}