CODEWITHSUNDEEP
pythonpandasplotseries
Eastsun
Eastsun

Reputation: 18859

How to make pandas's Series only plot the data in index?

EDIT The missed dates 20140103 and 20140104 is expected, I don't want them to be patched auto. And I don't want to use xs = pd.Series(data=range(len(ts)), index=pd.to_datetime(ts)) instead of xs = pd.Series(data=range(len(ts)), index=pd.to_datetime(ts)) since I want to use some operation as

xs = pd.Series(data=range(len(ts)), index=pd.to_datetime(ts))

xs['2014']
Out[24]: 
2014-01-01    0
2014-01-02    1
2014-01-05    2
2014-01-06    3
2014-01-07    4
dtype: int64

which doesn't work for:

In [26]: xs = pd.Series(data=range(len(ts)), index=ts)

In [27]: xs['2014']
---------------------------------------------------------------------------
KeyError: '2014'

For example:

In [1]: import pandas as pd

In [2]: ts = ['20140101', '20140102', '20140105', '20140106', '20140107']

In [3]: xs = pd.Series(data=range(len(ts)), index=pd.to_datetime(ts))

In [4]: xs.plot()

Will plot this image, which add the extra data on 20140103, 20140104. enter image description here

While I want to got a image like this:

import matplotlib.pyplot as plt

plt.plot(xs.values)

enter image description here

Upvotes: 1

Views: 2794

Answers (4)

Kirubaharan J
Kirubaharan J

Reputation: 2365

Thanks to euri10 for use_index = False

ts = ['20140101', '20140102', '20140105', '20140106', '20140107']
xs = pd.Series(data=range(len(ts)), index=pd.to_datetime(ts))
fig, ax = plt.subplots()
xs.plot(use_index=False)
ax.set_xticklabels(pd.to_datetime(ts))
ax.set_xticks(range(len(ts)))
fig.autofmt_xdate()
plt.show()

enter image description here

Upvotes: 3

Chitrasen
Chitrasen

Reputation: 1726

import pandas as pd
ts = ['20140101', '20140102', '20140105', '20140106', '20140107']

xs = pd.Series(data=range(len(ts)), index=pd.to_datetime(ts))
print xs


#output
2014-01-01    0
2014-01-02    1
2014-01-05    2
2014-01-06    3
2014-01-07    4

You are missing 2 dates.

Upvotes: 0

euri10
euri10

Reputation: 2626

Couldn't test it but you may want to use use_index=False and/or xticks

Upvotes: 2

brian
brian

Reputation: 131

Since you converted the dates to datetime, pandas is plotting the dates themselves on the x-axis. Since there are days that aren't included in the Series, pandas will leave the appropriate spaces on the x-axis where those points are. If, for some reason, you don't want this (for instance, if you are counting business days and the missing points are weekends), I think the solution that makes it most clear is to make the index not a date but a number (Day 1, Day 2, etc.). To my knowledge, pandas won't allow you to distort the axes.

Upvotes: 1

Related Questions