CODEWITHSUNDEEP
pythonpandas
Robs
Robs

Reputation: 5

Find times occurrences in python

I have a DataFrame from cvs.

I would like to know at what time there is greater probability to have the row of column "First" to 0 between 23:00 to 23:50.

                      Date First Second
0      2019-01-09 22:59:00     0     20
1      2019-01-09 23:04:00    14     32
2      2019-01-09 23:10:00     9     27
3      2019-01-09 23:11:00     7     27
4      2019-01-09 23:12:00     7     26
5      2019-01-09 23:13:00     7     26
6      2019-01-09 23:14:00     7     25
7      2019-01-09 23:15:00     6     25
8      2019-01-09 23:16:00     5     23
9      2019-01-09 23:17:00     4     22
10     2019-01-09 23:18:00     3     22
...                    ...   ...    ...
134761 2019-05-05 21:20:00    18     36
134762 2019-05-05 21:21:00    16     35
134763 2019-05-05 21:22:00    15     34
134764 2019-05-05 21:23:00    14     33

I use this code for select the time expected : 

heure = df.set_index('Date').between_time('23:00:00','23:50:00')

But I wasn't able to extract just time.

If you have any suggestions :)

Thanks,

Robin

Upvotes: 0

Views: 72

Answers (3)

ALollz
ALollz

Reputation: 59519

Filter based on time. Then find the most common time where First is 0.

try:
    (df.set_index('Date').between_time('23:00:00','23:50:00').reset_index()
       .loc[lambda x: x.First == 0].Date.dt.time.value_counts().index[0])
except IndexError:
    print('No matches')

This will return a datetime.time, or in the case of your sample data, it will print that nothing matches as there are no 0s between the specified times.

Upvotes: 1

smj
smj

Reputation: 1284

How about using dt accessors? Updated with an end to end example for your use case.

import pandas as pd
import numpy as np

df = pd.DataFrame(
    {
        'date': [
            pd.to_datetime('2019-01-09 22:59:00'),
            pd.to_datetime('2019-01-09 23:00:00'),
            pd.to_datetime('2019-01-09 23:49:59'),
            pd.to_datetime('2019-01-09 23:50:00'),
            pd.to_datetime('2019-01-09 23:51:00'),
        ],
        'value': [0, 0, 5, 6, 1]
    }        
)

# A mask to split the datset into two groups, based on the time.

df['in_range'] = np.where((df['date'].dt.hour == 23) & (df['date'].dt.minute < 50), 'In Range', 'Out of Range')

# A column that tests the condition you mentioned

df['condition'] = df['value'] == 0

# Group and get the average, which is the likelihood that value == 0, per group.

print(df.groupby('in_range')['condition'].mean())

Gives:

                    mask
In Range        0.500000
Out of Range    0.333333

Upvotes: 1

VietHTran
VietHTran

Reputation: 2318

You should convert your "Date" column data to datetime type first and you can apply the indexing method using dt as mentioned by @smj

import pandas as pd

df = pd.read_csv('./sample.csv')
df['Date'] = pd.to_datetime(df['Date'])
print df[(df['Date'].dt.hour == 23) & (df['Date'].dt.minute < 50)]

Upvotes: 0

Related Questions