问题
I have a data file that consists of several data points, that i would like to plot in a particular order. Here is the example data file:
cop 94.0528 313 441
cpr 225.172 726 747
ent 444.786 926 939
fsh 256.038 743 754
fsp 43.7764 340 356
pbp 343.708 825 834
rho 426.497 858 886
sca 342.431 427 872
I am currently plotting those just below each other in the way i set the example. How can i change the order of those data records within my python script to a specified order? I already tried to work this into an array. So i would do
data=Numpy.genfromtxt(txt)
transformdata.append(2) # cpr to the first slot
transformdata.append(1) # cop to the second slot
outputarray.append(data[transformdata(1)]
outputarray.append(data[transformdata(2)]
pos = range(1,size(outputarray))
scatter(outputarray, pos)
However this is messy and certainly not the best way to accomplish this. How would i do this using Numpy or Matplotlib libraries?
回答1:
Assuming that transformdata is a ndarray, like data, you could use a trick of indexing to get a view of your array without having to copy it. For example:
>>> x=np.array(zip("abcdef",range(5),np.random.rand(5)),dtype=[('',"|S1"),('',int),('',float)])
>>> print x
[('a', 0, 0.9818028495949546) ('b', 1, 0.08931155317859962)
('c', 2, 0.018796963462246796) ('d', 3, 0.275603618214155)
('e', 4, 0.6328806434997251)]
>>> y = x[[1,3,0,2,4]]
>>> print y
[('b', 1, 0.08931155317859962) ('d', 3, 0.275603618214155)
('a', 0, 0.9818028495949546) ('c', 2, 0.018796963462246796)
('e', 4, 0.6328806434997251)]
The key here is x[[1,3,0,2,4]], where you're indexing with a list of the indices you want in the order you need.
来源：https://stackoverflow.com/questions/12035603/how-do-i-control-the-sequence-of-records-in-a-bar-plot-with-matplotlib