Eroman: Created page with "= Lecture 5 = Working with real data.
Datetime, CSV files == Datetime == import datetime as dt my_datetime = dt.datetime.utcnow() # get the cu..."

2017-05-04T08:21:00Z

Created page with "= Lecture 5 = Working with real data. Datetime, CSV files == Datetime == <source lang="python">import datetime as dt my_datetime = dt.datetime.utcnow() # get the cu..."

New page

= Lecture 5 =

Working with real data. 
Datetime, CSV files

== Datetime ==

<source lang="python">import datetime as dt

my_datetime = dt.datetime.utcnow() # get the current date and time (without timezone) as a datetime object

one_day = dt.timedelta(days=1) # create a duration of 1 day

tomorrow = my_datetime + one_day

yesterday = my_datetime - one_day

tomorrow > yesterday # True

current_year = my_datetime.year

first_day_of_year = dt.datetime(year=current_year, month=1, day=1)
</source>
== Task 1 ==

Display the current date like 
“2017-05-04”

Calculate How many days since Jan 1.

Optional, requires using timezone: 
How many hours are left until midnight today?

<source lang="python">import datetime as dt

print(dt.datetime.utcnow().strftime("%Y-%m-%d"))

today = dt.date.today()

first_of_the_year = dt.date(year=2017, month=1, day=1)

print(today - first_of_the_year)

now = dt.datetime.now()
midnight = dt.datetime(year=now.year, month=now.month, day=(now.day + 1))
print(midnight - now)</source>
== CSV files ==

CSV files are a portable and common way of using tabular data. They are easily used in python.

== Example: Printing every row of a csv file ==

Using a hardcoded path to <code>~/sheet1.csv</code>

<source lang="python">"""
Uses a hardcoded path to the file. used when you cannot use the command line to run python. For example if you run the script in an IDE
"""
import csv
from os.path import expanduser, join

home = expanduser("~")
path = join(home, 'sheet1.csv')
csv_file = open(path)

reader = csv.reader(csv_file)
for row in reader:
print(line)

csv_file.close()</source>
Using the argparse library

<source lang="python">"""
Open a file using argparse.
Work when ran from command line
"""
import argparse
import csv

parser = argparse.ArgumentParser()
parser.add_argument('file')
args = parser.parse_args()

csv_file = open(args.file)

reader = csv.reader(csv_file)
for row in reader:
print(row)

csv_file.close()</source>
== Task 2 ==

What was the average price of bitcoin in April?

CSV of historical price data here: 
https://blockchain.info/charts/market-price?timespan=1year

=== Hints ===

convert timestamps to datetime object

<pre>datetime.datetime.strptime("2016-05-11 00:00:00", "%Y-%m-%d 00:00:00")</pre>
=== Example solution ===

<source lang="python">"""
Open a file using argparse.
Work when ran from command line
"""
import datetime as dt
import argparse
import csv

MIN_DATE = dt.datetime(year=2017, month=4, day=1)
MAX_DATE = dt.datetime(year=2017, month=5, day=1)

parser = argparse.ArgumentParser()
parser.add_argument('file')
args = parser.parse_args()
csv_file = open(args.file)
reader = csv.reader(csv_file)

prices_in_april = []

for row in reader:
raw_price = row[1]
price = float(raw_price)
raw_date = row[0]
date = dt.datetime.strptime(raw_date, "%Y-%m-%d 00:00:00")
if date >= MIN_DATE and date < MAX_DATE:
prices_in_april.append(price)

csv_file.close()

print(sum(prices_in_april) / len(prices_in_april))</source>

I719 Fundamentals of Python/lecture5v2 - Revision history

Eroman: Created page with "= Lecture 5 = Working with real data. Datetime, CSV files == Datetime == import datetime as dt my_datetime = dt.datetime.utcnow() # get the cu..."

Eroman: Created page with "= Lecture 5 = Working with real data.
Datetime, CSV files == Datetime == import datetime as dt my_datetime = dt.datetime.utcnow() # get the cu..."