how to remove positive infinity from numpy array...if it is already converted to a number?

First of all, 1.79769313486e+308 is not the same as +inf. The former is the largest number which can be expressed with a 64-bit float, the latter is a special float.

If you just have very large numbers in your array, then:

A[A > 1e308] = 0

is sufficient. Thet'll replace oll elements above 1e308 with 0.

It is also possible to operate with the inf's. For example:

>>> fmax = np.finfo(np.float64).max
>>> pinf = float('+inf')
>>> ninf = float('-inf')
>>> fnan = float('nan')
>>> print fmax, pinf, ninf, fnan
1.79769313486e+308 inf -inf nan

So, these are completely different things. You may compare some of them:

>>> pinf > fmax
True
>>> ninf < 0.0
True
>>> pinf == pinf
True
>>> pinf == ninf
False

This looks good! However, nan acts differently:

>>> fnan > 0
False
>>> fnan < 0
False
>>> fnan == 0
False
>>> fnan < pinf
False
>>> fnan == fnan
False

You may use positive and negativi infinities with Numpy ndarraywithout any problems. This will work:

A[A == pinf] = 0.0

But if you have nans in the array, you'll get some complaints:

>>> np.array([fnan, pinf, ninf]) < 0
RuntimeWarning: invalid value encountered in less
[False, False, True]

So, it works but complains => do not use. The same without the nan:

>>> np.array([0.0, pinf, ninf]) < 0
[False, False, True]

If you want to do something with the nans (should you have them), use numpy.isnan:

A[np.isnan(A)] = 0.0

will change all nans into zeros.

---

And -- this you did not ask -- here is one to surprise your friends (*):

>>> [float('-0.0'), 0.0] * 3
[-0.0, 0.0, -0.0, 0.0, -0.0, 0.0]

Yep, float64 (and float32) have even a separate -0.0. In calculations it acts as an ordinary zero, though:

>>> float('-0.0') == 0.0
True

(*) Depending on the kind of people you call friends.

To remove the very high values:

>>> a = numpy.array([1, 2, 1.8E308, 1.8E308, 42])
>>> a[a < 1E308] # use whatever threshold you like
array([  1.,   2.,  42.])

To set them 0:

>>> a = numpy.array([1, 2, 1.8E308, 1.8E308, 42])
>>> a[a >= 1E308] = 0
>>> a
array([  1.,   2.,   0.,   0.,  42.])