如何将小时分钟和秒数放在数据框中而不是年，月和日？

Question

I want to plot a line graph of ECG in mV and time in HH:MM:SS:MMM . its a 10 second ECG strip.

1. image of ECG CSV file with two ECG values and time

I Have extract the Time column and now i want to convert the time column in dataframe of python and then plot it on graph

but when I apply to_datetime() function it give me the following error

to assemble mappings requires at least that [year, month, day] be specified: [day,month,year] is missing

2. Screenshot of error i get

please Help me to resolve this error , I only want to put %H:%M:%S.%f because i don not have the year , months and days.

Answer 1

See to_timedelta

Date :

df1 = {'Time':['11:20:15.333','10:00:00.444'],'P1':['102','102'],'P2':['240','247']}
df1 = pd.DataFrame(data=df1)
df1

Code :

df1['Time'] = pd.to_timedelta(df1['Time'])
df1

Result :

    Time            P1  P2
0   11:20:15.333000 102 240
1   10:00:00.444000 102 247

Reference : https://stackoverflow.com/a/46801500/1855988

Answer 2

You need to specify the format you want to convert the time to. You can find out more information here about what each symbol means.

# before it is object
df['column_name'] = pd.to_datetime(df['column_name'], format="%H:%M:%S,%f")
df = df.astype('datetime64')
df['column_name'] =  pd.to_datetime(df['column_name'], format='%H:%M:%S', errors='coerce').dt.time
# coming back to object
print(df.head())
# print(df.info())

Answer 3

As commented you can add a date to those times. The date can be arbitrary. Then you can convert to datetime and use them to plot your graph.

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates

data = {'Time':['11:20:15.333','12:00:00.444', '13:46:00.100'],
       'A':[1,3,2],'B':[5,5,4]}
df = pd.DataFrame(data=data)
df["Date"] = "2019-09-09"
df['Datetime'] = pd.to_datetime(df['Date']) + pd.to_timedelta(df['Time'])

df = df[["Datetime", "A", "B"]].set_index("Datetime")

ax = df.plot(x_compat=True)
ax.xaxis.set_major_formatter(mdates.DateFormatter("%H:%M:%S.%f"))

plt.show()