Python: plot timestamp data frame matplotlib

Question

I am plotting values over the time (i.e. hourly) I have a dataframe dfout as following

    val     x            y                      ID                              time
0   6.0 48.730304   11.594837   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:02:02
1   6.5 48.731639   11.602004   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:02:26
2   7.0 48.740104   11.641433   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:04:45
3   6.5 48.744026   11.648048   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:05:11
4   6.0 48.747356   11.654539   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:05:32
5   5.5 48.749050   11.660844   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:05:50
6   5.0 48.749935   11.666812   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:06:10
7   4.5 48.751007   11.677590   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:06:54
8   4.0 48.742317   11.675558   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:08:31
9   4.5 48.736461   11.675782   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:08:55
10  5.0 48.729659   11.675481   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:09:20
11  5.5 48.725871   11.673145   0efd2316-feb9-461c-b102-045aef8e22fe    2014-03-10 08:09:35

and I want to plot the column val as following

import matplotlib.dates as md
fig=figure(figsize=(20,10))
ax=plt.gca()
xfmt = md.DateFormatter('%H:%M:%S')
plt.xticks( rotation=25 )
ax.xaxis.set_major_formatter(xfmt)
plt.title("Temperature")
plt.ylabel(r'Temperature, $^{o}C$')
ax.plot(datenumes,dfout.val, marker='.',lw=0)

However I would like the xlabel between 08:00:00 to 16:00:00 hour by hour, so: 08:00:00, 09:00:00, 10:00:00,....,16:00:00.

Answer 1

Simply choose the position of your ticks using ax.set_xticks() and feed it a custom list of datetime objects created for that purpose. The issue is that the label representations of those object are floats. You can remedy this by creating another list that contains the time-strings as you want them to look like, which I called ticklabels in this solution.

import matplotlib.dates as md
import matplotlib.pyplot as plt
from  datetime import *

val = [6, 6.5, 7, 6.5, 6, 5.5, 5, 4.5, 4, 4.5, 5, 5.5]
datenumes = [datetime(2014, 3, 10, 8, 2, 2), 
    datetime(2014, 3, 10, 8, 2, 26), 
    datetime(2014, 3, 10, 10, 4, 45), 
    datetime(2014, 3, 10, 10, 5, 11), 
    datetime(2014, 3, 10, 12, 5, 32),
    datetime(2014, 3, 10, 12, 5, 50), 
    datetime(2014, 3, 10, 15, 6, 10), 
    datetime(2014, 3, 10, 16, 6, 54), 
    datetime(2014, 3, 10, 17, 8, 31), 
    datetime(2014, 3, 10, 18, 8, 55), 
    datetime(2014, 3, 10, 19, 9, 20), 
    datetime(2014, 3, 10, 22, 9, 35)]

# creates list of datetimes where ticks should be set
ticks = [datetime(2014, 3, 10, 8, 0, 0), 
        datetime(2014, 3, 10, 10, 0, 0),
        datetime(2014, 3, 10, 12, 0, 0),
        datetime(2014, 3, 10, 14, 0, 0),
        datetime(2014, 3, 10, 16, 0, 0),
        datetime(2014, 3, 10, 18, 0, 0),
        datetime(2014, 3, 10, 20, 0, 0),
        datetime(2014, 3, 10, 22, 0, 0) ]

# generate labels for the datetimes used as ticks
ticklabels = [str(tick)[11:] for tick in ticks]

#  PLOT
fig=plt.figure(figsize=(20,10))
ax=plt.gca()
plt.xticks( rotation=25 )

# set the ticks according to the list tick
ax.set_xticks(ticks)
##  set labels to human readable format
ax.set_xticklabels(ticklabels)

plt.title("Temperature")
plt.ylabel(r'Temperature, $^{o}C$')
ax.plot(datenumes, val, marker='.', lw=0)
plt.show()