Iterating over time series

Question

The following is a sample of my dataset:

index      time      speed
 0      00:00:00      15
 1      00:00:05      18
 2      00:00:10      23
 3      00:00:15      25
 4      00:00:20      34

I would like to create a for loop that does the same function as below:

for i in range (0,5,1):
    if df.speed[i] > df.speed [i+2]:
         print ('Larger')
    else:
         print('Smaller')

However, I would like to refer to time instead of indices in the FOR loop. For example:

for t in range (00:00:00, 00:00:20 , 5s):
  if df.speed[t] > df.speed [t+10s]: 
     print ('Larger')
  else:
     print('Smaller')

So the FOR LOOP will take the speed value at a certain t and compare with it the value of the speed after 10 seconds. If it is larger, then it prints Larger, otherwise, Smaller.

I appreciate any help. Thanks.

Answer 1

Date Range (Seconds)

It seems you're using pandas. If that's so, you really should look at it's Time Series features: https://pandas.pydata.org/pandas-docs/stable/timeseries.html

More specific, at pandas.date_range():

import pandas as pd

pd.date_range('00:00:00', '00:00:20', freq='5s')

Since it's a date range function, it'll return datetimes

DatetimeIndex(['2019-03-06 00:00:00',
               '2019-03-06 00:00:05',
                ...,
               '2019-03-06 00:00:20'], dtype='datetime64[ns]', freq='5S')

From here just use .strftime() to get what you want:

pd.date_range('00:00:00', '00:00:20', freq='5s').strftime('%H:%M:%S')

Returns

Index(['00:00:00', '00:00:05', ..., '00:00:20'], dtype='object')

---

Access dataframe by time

After that, in order to access your dataframe using this index you have two options:

1. Using .loc to find the row where this time is:

for time in pd.date_range('00:00:00', '00:00:20', freq='5s').strftime('%H:%M:%S'):
    if (df.loc[df['time'] == time, 'speed'] ..):

2. By redefining the dataframe index with set_index(), and then accessing it directly by time:

df.set_index('time', inplace=True)
for time in pd.date_range('00:00:00', '00:00:20', freq='5s').strftime('%H:%M:%S'):
    if (df.speed[time] ...):

---

Edit to address comment question

After you you use strftime() it is converted to a String. You could convert it back to add the 10 seconds like:

df.speed[(pd.to_datetime(time) + pd.to_timedelta(10, unit='s')).strftime('%H:%M:%S')].

Or:

for time in pd.date_range('00:00:00', '00:00:20', freq='5s'):
    if (df.speed[time.strftime('%H:%M:%S')]>df.speed[(time + pd.to_timedelta(10, unit='s')).strftime('%H:%M:%S')]):