3 ways to parse time strings in python [Note]
What you want to do
I want to make a memorandum about how to parse the time written in text to the datetime type of python.
environment
- Windows10 pro
- anaconda navigator 1.6.2
- Jupyter Notebook 5.0.0
- python 3.6.1
- Intel Core i7 4770 16GB
Background
-
- The date and time appearing in the XML of OpenWeatherMap was in a format I had never seen (only for myself) and the time was GMT (UTC). 2.9 It is difficult to deal with the situation where the dates are different due to the time difference unless the process of adding 9.9 hours is performed.
-
- I thought it would be best to put it in a pre-prepared mold and calculate the time difference additionally.
By the way, the date that appears in XML looks like this.
2017-07-08T19:33:11
I found a lot when I looked up the method of parsing, so I will summarize the code.
import time
from datetime import datetime
from dateutil.parser import parse
start = time.time()
date="2017-07-08T19:33:11" #Character string to be parsed
for i in range(100000): #parse()Loop to parse 100,000 times
dt1 = parse(date)
elapsed_time_for_parse = time.time() - start #parse()Stores the elapsed time when parsing with
start = time.time()
for i in range(100000): #strptime()Loop to parse 100,000 times
dt2 = datetime.strptime(date, '%Y-%m-%dT%H:%M:%S')
elapsed_time_for_strptime = time.time() - start #strptime()Stores the elapsed time when parsing with
start = time.time()
for i in range(100000): #Loop that parses 100,000 times by cutting out a character string
dt3 = datetime(
year=int(date[0:4]),month=int(date[5:7]),day=int(date[8:10]),
hour=int(date[11:13]),minute=int(date[14:16]),second=int(date[17:20])
)
elapsed_time_for_cutparse = time.time() - start #Stores the elapsed time when parsing with character string cutout
print('parse() takes {:.4f}sec, dt1:{}'.format(elapsed_time_for_parse,dt1))
print('strptime() takes {:.4f}sec, dt2:{}'.format(elapsed_time_for_strptime,dt2))
print('cut and parse takes {:.4f}sec, dt3:{}'.format(elapsed_time_for_cutparse,dt3))
result
| Method | Time required |
|---|---|
| parse() | 5.8433 |
| strptime() | 1.0622 |
| String cutout | 0.2875 |
- The time required is the time when the loop was executed 100,000 times in the above environment of the author.
Summary
parse () is easy but slow with one shot It takes a lot of time to cut out character strings, but it is fast.
The site that I used as a reference
OpenWeatherMap Code 7 ward Date and time strings are parsed with python. Extract the year, month and hour miyalog Python datetime.strptime is very slow
result screen of jupyter notebook
Recommended Posts