[Tutor] Loops and matrices'

[Peter Otten]

Reed DA (Danny) at Aera wrote:

> I have a matrix which contains temperatures. The columns are time, spaced
> 33 seconds apart, and the rows are depth intervals. What I'm trying to do
> is create another matrix that contains the rate of change of temperature
> for each depth. The array is called LS_JULY_11 and the output array should
> be delta_temp

I think you don't need any explicit for-loops:

>>> import numpy as np
>>> july = np.random.randint(0, 100, (3, 4))
>>> july
array([[27, 43, 67, 12],
       [52, 22, 54, 26],
       [70, 81, 61, 49]])
>>> dt = 33.0
>>> july[1:]
array([[52, 22, 54, 26],
       [70, 81, 61, 49]])
>>> july[:-1]
array([[27, 43, 67, 12],
       [52, 22, 54, 26]])
>>> (july[1:]-july[:-1])
array([[ 25, -21, -13,  14],
       [ 18,  59,   7,  23]])
>>> (july[1:]-july[:-1])/dt
array([[ 0.75757576, -0.63636364, -0.39393939,  0.42424242],
       [ 0.54545455,  1.78787879,  0.21212121,  0.6969697 ]])

If rows and columns are swapped in the above, just transpose the matrix 
before you start and again when you're done:

>>> july = july.transpose()
>>> july
array([[27, 52, 70],
       [43, 22, 81],
       [67, 54, 61],
       [12, 26, 49]])
>>> ((july[1:]-july[:-1])/dt).transpose()
array([[ 0.48484848,  0.72727273, -1.66666667],
       [-0.90909091,  0.96969697, -0.84848485],
       [ 0.33333333, -0.60606061, -0.36363636]])