I have an array populated with the same value, however, I want this value to incrementally decrease by a value of 0.025 with each row. It currently looks like this:
import numpy as np
vp_ref = 30
surf_lay = np.ones(([1000,10000]), dtype=np.float32);
gvp_ref = vp_ref * surf_lay
So the array is filled with 30s. I want the first row to be 30, decrease to 29.975 in the next row, and keep going until the bottom. How would I do this?
Here's a solution:
step_range to get all the values starting from 0, adding step until the end of your matrix size.step = 0.025
step_range = np.arange(0, gvp_ref.shape[0] * step, step).reshape(-1, 1)
print(gvp_ref - step_range)
Output:
array([[30. , 30. , 30. , ..., 30. , 30. , 30. ],
[29.975, 29.975, 29.975, ..., 29.975, 29.975, 29.975],
[29.95 , 29.95 , 29.95 , ..., 29.95 , 29.95 , 29.95 ],
...,
[ 5.075, 5.075, 5.075, ..., 5.075, 5.075, 5.075],
[ 5.05 , 5.05 , 5.05 , ..., 5.05 , 5.05 , 5.05 ],
[ 5.025, 5.025, 5.025, ..., 5.025, 5.025, 5.025]])
np.full broadcasting is a bit of a stretch. You could use np.broadcast_to and then it would be true. Actually, I'm pretty sure broadcasting would work as gvp_ref - step_range...
gvp_ref - step_range would work but it didn't initially. In fact, with .reshape(1000,1), you're right, it does! Thanks @AndrasDeak, edited, that's cleaner.You can use np.linspace for creating the linearly spaced data and then np.tile for creating the 2D array:
n = 1000
tmp = np.linspace(30, 30 - (n-1)*0.025, n)
result = np.tile(tmp[:, None], (1, 10_000))
Here, in this code alpha is 0.025
import numpy as np
vp_ref = 30
surf_lay = np.ones(([1000,10000]), dtype=np.float32);
gvp_ref = vp_ref * surf_lay
alpha = 0.025
substration_array = np.array([[alpha*i]*gvp_ref.shape[1] for i in range(gvp_ref.shape[0])])
gvp_ref.shape
substration_array.shape
output = np.subtract(gvp_ref, substration_array)
