Creating array with values less than the threshold using Python-CodePudding

I want to pick the minimum in the first row and then select values which are less than this minimum as I scan the succeeding rows. The output should look as below.

import numpy as np
a = np.array([[0.4234101 , 0.73587042, 0.81, 0.83202077, 0.73592897],
       [0.45714687, 0.13144797, 0.67110076, 0.70283126, 0.32859529],
       [0.30888356, 0.84115786, 0.24648344, 0.4963486 , 0.42780253],
       [0.10956774, 0.49696239, 0.17086982, 0.34758674, 0.6332388 ],
       [0.69931352, 0.72449178, 0.98831224, 0.20775389, 0.19041985]])

print(a<0.4234101)

The desired output is

array([0.4234101 , 0.13144797, 0.32859529, 0.30888356, 0.24648344,
       0.10956774, 0.17086982, 0.34758674, 0.20775389, 0.19041985])

CodePudding user response：

You could index the first row, find its minimum and create a boolean mask to filter a with.

a[a < a[0].min()]

If you want the minimum as part of the final output, then you could concatenate it to the output above:

first_row_min = a[0].min()
np.r_[first_row_min, a[a < first_row_min]]

Output:

array([0.4234101 , 0.13144797, 0.32859529, 0.30888356, 0.24648344,
       0.10956774, 0.17086982, 0.34758674, 0.20775389, 0.19041985])

CodePudding user response：

Try:

print([i  for i in a.reshape(-1,) if i<0.42])

or if you want it linked to the first value:

print([i  for i in a.reshape(-1,) if i<a.reshape(-1,)[0]])

CodePudding user response：

One of possible solutions is:

Get the minimum from row 0:
```
a0min = a[0].min()
```
Get following rows:
```
a1 = a[1:]
```
Get the result:
```
result = np.insert(a1[a1 < a0min], 0, a0min)
```
i.e. get the wanted elements from a1 (following rows) and insert the minimum from the initial row at the beginning.

The result is:

array([0.4234101 , 0.13144797, 0.32859529, 0.30888356, 0.24648344,
       0.10956774, 0.17086982, 0.34758674, 0.20775389, 0.19041985])

You don't need any explicit iteration over the source array, as it is performed under the hood by Numpy.

CodePudding user response：

The method you might want is np.nditer() which iterates over the single elements of an multidimensional array

import numpy as np

a = np.array([[0.4234101 , 0.73587042, 0.81, 0.83202077, 0.73592897],
       [0.45714687, 0.13144797, 0.67110076, 0.70283126, 0.32859529],
       [0.30888356, 0.84115786, 0.24648344, 0.4963486 , 0.42780253],
       [0.10956774, 0.49696239, 0.17086982, 0.34758674, 0.6332388 ],
       [0.69931352, 0.72449178, 0.98831224, 0.20775389, 0.19041985]])

values = []
min = a[0].min()

for element in np.nditer(a[1:]):
    if element < min:
        values.append(element.item(0))