Plotting dates with matplotlib

Question

I've been attempting to plot data from a comma delimited csv file which contains a date and a float:

Date,Price (€)
01062017,20.90
02062017,30.90
03062017,40.90
04062017,60.90
05062017,50.90

I then attempt to plot this with the following code:

import matplotlib.pyplot as plt
import numpy as np
import datetime

dates,cost = np.loadtxt('price_check.csv',delimiter=',',skiprows=1,unpack=True)

xdates = [datetime.datetime.strptime(str(int(date)),'%d%m%Y') for date in dates]

fig = plt.figure()
ax = plt.subplot(111)
plt.plot(xdates, cost,'o-',label='Cost')
plt.legend(loc=4)
plt.ylabel('Price (Euro)')
plt.xlabel('date')
plt.gcf().autofmt_xdate()
plt.grid()
plt.savefig('sunglasses_cost.png')
plt.show()

However, when the data is plotted, it looks like the leading zero in in the date string is being dropped:

Is there an easy way for the full date to be used in the plot?

Why are you turning your date to integer before giving it to datetime.datetime.strptime()? That's what kills your leading zeros. — zwer
– zwer, Commented Jun 4, 2017 at 20:46

ImportanceOfBeingErnest · Accepted Answer · 2017-06-04 21:22:41Z

The problem are the dates, which are converted to integers and loose their leading zero. Then "01062017" becomes 1062017 and is then interpreted as (2017, 6, 10, 0, 0), so 2 digits as day, one digit month. For 5062017, because there is no 50th of june, it is interpreted differently and correctly as (2017, 6, 5, 0, 0).

The least invasive method to overcome this would be to format the string such that it always has 8 digist before datetime conversion:

xdates = [datetime.datetime.strptime('{:08}'.format(int(date)),'%d%m%Y') for date in dates]

This will then result in the correct plot. However, the xticklabels may show in an inconvenient way. This could be adjusted by choosing some locator and formatter

import matplotlib.dates as mdates
plt.gca().xaxis.set_major_locator(mdates.DayLocator())
plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%d-%m-%Y'))

As a last comment: If you have the choice to select the format of your input file, it might be worth specifing it in a non-ambiguous way, e.g. 20170601.

Collectives™ on Stack Overflow

Plotting dates with matplotlib

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related