how can i delete whole day rows on condition column values.. pandas

问题

i have below times series data frames

i wanna delete rows on condtion (check everyday) : check aaa>100 then delete all day rows (in belows, delete all 2015-12-01 rows because aaa column last 3 have 1000 value)

     ....
   date       time    aaa
2015-12-01,00:00:00,0
2015-12-01,00:15:00,0
2015-12-01,00:30:00,0
2015-12-01,00:45:00,0
2015-12-01,01:00:00,0
2015-12-01,01:15:00,0
2015-12-01,01:30:00,0
2015-12-01,01:45:00,0
2015-12-01,02:00:00,0
2015-12-01,02:15:00,0
2015-12-01,02:30:00,0
2015-12-01,02:45:00,0
2015-12-01,03:00:00,0
2015-12-01,03:15:00,0
2015-12-01,03:30:00,0
2015-12-01,03:45:00,0
2015-12-01,04:00:00,0
2015-12-01,04:15:00,0
2015-12-01,04:30:00,0
2015-12-01,04:45:00,0
2015-12-01,05:00:00,0
2015-12-01,05:15:00,0
2015-12-01,05:30:00,0
2015-12-01,05:45:00,0
2015-12-01,06:00:00,0
2015-12-01,06:15:00,0
2015-12-01,06:30:00,1000
2015-12-01,06:45:00,1000
2015-12-01,07:00:00,1000
         ....

how can i do it ?

回答1:

I think you need if MultiIndex first compare values of aaa by condition and then filter all values in first level by boolean indexing, last filter again by isin with inverted condition by ~:

print (df)
                      aaa
date       time          
2015-12-01 00:00:00     0
           00:15:00     0
           00:30:00     0
           00:45:00     0
2015-12-02 05:00:00     0
           05:15:00   200
           05:30:00     0
           05:45:00     0
2015-12-03 06:00:00     0
           06:15:00     0
           06:30:00  1000
           06:45:00  1000
           07:00:00  1000

lvl0 = df.index.get_level_values(0)
idx = lvl0[df['aaa'].gt(100)].unique()
print (idx)
Index(['2015-12-02', '2015-12-03'], dtype='object', name='date')

df = df[~lvl0.isin(idx)]
print (df)
                     aaa
date       time         
2015-12-01 00:00:00    0
           00:15:00    0
           00:30:00    0
           00:45:00    0

And if first column is not index only compare column date:

print (df)
          date      time   aaa
0   2015-12-01  00:00:00     0
1   2015-12-01  00:15:00     0
2   2015-12-01  00:30:00     0
3   2015-12-01  00:45:00     0
4   2015-12-02  05:00:00     0
5   2015-12-02  05:15:00   200
6   2015-12-02  05:30:00     0
7   2015-12-02  05:45:00     0
8   2015-12-03  06:00:00     0
9   2015-12-03  06:15:00     0
10  2015-12-03  06:30:00  1000
11  2015-12-03  06:45:00  1000
12  2015-12-03  07:00:00  1000

idx = df.loc[df['aaa'].gt(100), 'date'].unique()
print (idx)
['2015-12-02' '2015-12-03']

df = df[~df['date'].isin(idx)]
print (df)
         date      time  aaa
0  2015-12-01  00:00:00    0
1  2015-12-01  00:15:00    0
2  2015-12-01  00:30:00    0
3  2015-12-01  00:45:00    0

来源：https://stackoverflow.com/questions/48556510/how-can-i-delete-whole-day-rows-on-condition-column-values-pandas

标签

python

pandas

time-series

row