在Python中：给定一个原始的数字数组，我如何创建一个包含特定范围内原始数组值的新数组？

Question

import os
import pyfits as ps
import lomb
import numpy as np
import matplotlib.pyplot as plt

hdulist = ps.open('filename')

tbdata = hdulist[1].data

PDCFlux = tbdata.field(7)
PDCFlux = PDCFlux[~np.isnan(PDCFlux)]

psd,freq = lomb.lomb(Timesec,PDCFlux)
logpsd = np.log10(psd)
logfreq = np.log10(freq)

#want an associated 'slope'.

'''PDC Slope'''
PDCFluxrange = PDCFlux.all[np.log10(4*(10**(-5)))<freq<np.log10(4*(10**(-7)))]

In the last line, I attempted to create a new array: PDCFluxrange that consists of values of the array PDCFlux for freq (frequency) between: log10(4*(10**(-5))) and log10(4*(10**(-7)) .

When I run the program in Pylab, I receive an error message similar to: Value Error: The truth value of an array with more than one element is ambiguous . Use a.any() or a.all() .

In which case, I am unsure of how to properly use either of these two functions in my code without creating a loop, or defining a new function; I think that there may be a simpler method to accomplish the same thing.

Answer 1

>>> import numpy as np
>>> a = np.arange(10)
>>> a[(a > 2) & (a < 8)]
array([3, 4, 5, 6, 7])

Each condition ( a > 2 and a < 8 ) is creating a boolean numpy array which is equal in length to a , and contains the truth value of your criterion at each position in a .

   >>> a > 2
   array([False, False, False,  True,  True,  True,  True,  True,  True,  True], dtype=bool)

Then, you're using numpy.logical_and to find the positions at which both conditions are true.

Finally, you're using the resulting array as a boolean or "mask" index array to retrieve only the elements from the original array that meet your criteria.

I think you're getting the ValueError because while a > 2 is valid (and is really just calling a.__gt__(2) , a statement like 2 < a < 8 is first calling cmp(2,a) , which results in the same ValueError for me:

>>> a = np.arange(10)
>>> cmp(2,a)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Answer 2

I think you want

PDCFluxrange = [x for x in PDCFlux if x>np.log10(4*(10**(-5))) and x<np.log10(4*(10**(-7)))]

Does that work for you? I hope I'm understanding your code right.