时间戳

首页 > 解决方案 > 计算时间间隔内列值的平均值

python - 计算时间间隔内列值的平均值

问题描述

我有数据框

                        id      timestamp               data    gradient        Start
timestamp                                       
2020-01-15 06:12:49.213 40250   2020-01-15 06:12:49.213 20.0    0.00373         NaN 
2020-01-15 06:12:49.313 40251   2020-01-15 06:12:49.313 19.5    0.00354         0.0 
2020-01-15 08:05:10.083 40256   2020-01-15 08:05:10.083 20.0    0.00020         1.0 
2020-01-15 08:05:10.183 40257   2020-01-15 08:05:10.183 20.5    -0.00440        0.0
                            ...
2020-01-31 09:01:50.993 40310   2020-01-31 09:01:50.993 21.0    0.55473         1.0
2020-01-31 09:01:51.093 40311   2020-01-31 09:01:51.093 21.5    0.00589         0.0
                            ...

我想找到data介于两者之间start_time ==1的平均值30 seconds。

可重现的例子：

d = {'timestamp':["2020-01-15 06:12:49.213", "2020-01-15 06:12:49.313", "2020-01-15 08:05:10.083", "2020-01-15 08:05:10.183", "2020-01-15 09:01:50.993", "2020-01-15 09:01:51.093", "2020-01-15 09:51:01.890", "2020-01-15 09:51:01.990", "2020-01-15 10:40:59.657", "2020-01-15 10:40:59.757", "2020-01-15 10:42:55.693", "2020-01-15 10:42:55.793", "2020-01-15 10:45:35.767", "2020-01-15 10:45:35.867", "2020-01-15 10:45:46.770", "2020-01-15 10:45:46.870", "2020-01-15 10:47:19.783", "2020-01-15 10:47:19.883", "2020-01-15 10:47:22.787"],
'data': [20.0, 19.5, 20.0, 20.5, 21.0, 21.5, 22.0, 22.5, 23.0, 23.5, 23.0, 22.5, 23.0, 23.5, 24.0, 24.5, 25.0, 25.5, 26], 
'gradient': [NaN, NaN, 0.000000, 0.000148, 0.000294, 0.000294, 0.000339, 0.000339, 0.000334, 0.000334, 0.000000, -0.008618, 0.000000, 0.006247, 0.090884, 0.090884, 0.010751, 0.010751, 0.332889],
'Start': [0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,]
}

df = pd.DataFrame(d)

预期输出：

start_time               end_time                   Average
2020-01-15 08:05:10.083  2020-01-15 09:01:51.093    20.25  = average of (20.0, 20.5)
2020-01-15 10:45:35.767  2020-01-15 10:45:35.767    23.75  = average of (23.0, 23.5, 24.0, 24.5)

编辑：

使用@jezrael 的代码：

df['timestamp'] = pd.to_datetime(df['timestamp'])
df['g'] = df['Start'].cumsum()

df1 = df[df['g'].ne(0)].copy()
#
s = df1.groupby('g')['timestamp'].transform('first')
df1 = df1[df1['timestamp'].between(s, s + pd.Timedelta(30, 's'))]
#
df2 = df1.groupby('g').agg(start_time=('timestamp','first'),
                           end_time=('timestamp','last'),
                           Average=('data','mean')).reset_index(drop=True)
print (df2)

我得到了输出 在此处输入图像描述

似乎有些开始和结束时间非常接近，大约相差 0.1 秒。这是数据采集设备的故障，每次记录 2 个数据点，而不是 1 个，并且数据点0.5有data. 此外，数据点很少，导致开始和结束时间在一个30 seconds时间间隔内非常接近。我的问题是，如果我们向前填充样品，是否有可能？以便有更多的数据来衡量。

标签： pythonpandasnumpy

解决方案

首先获取timestamp每组GroupBy.transform，GroupBy.first然后比较Series.between：

df['timestamp'] = pd.to_datetime(df['timestamp'])
df['g'] = df['Start'].cumsum()

df1 = df[df['g'].ne(0)].copy()
#
s = df1.groupby('g')['timestamp'].transform('first')
df1 = df1[df1['timestamp'].between(s, s + pd.Timedelta(30, 's'))]
#
df2 = df1.groupby('g').agg(start_time=('timestamp','first'),
                           end_time=('timestamp','last'),
                           Average=('data','mean')).reset_index(drop=True)
print (df2)
               start_time                end_time  Average
0 2020-01-15 08:05:10.083 2020-01-15 08:05:10.183    20.25
1 2020-01-15 10:45:35.767 2020-01-15 10:45:46.870    23.75

推荐阅读