If you don’t know the full size of your data but need to add data incrementally, you can add rows to the array to accumulate these results. This allows you to build an array structure dynamically.

The easiest and most straightforward way to append a new row to an empty numpy array is to use the “numpy.append()” method. It accepts an empty array and the array to be appended and returns the new array with the appended row.

Here is the step-by-step guide:

Step 1: Import numpy

Before importing numpy, we need to install the library using the below command:

pip install numpy

Now, import the numpy library:

import numpy as np

Step 2: Creating an empty array

You can use the np.array([]) snippet to create an empty numpy array. This will be the array to which we append a new row.

empty_array = np.array([])

Step 3: Create the row (array) you want to append

Let’s define a row (an array with values) that will be appended to empty_array.

row_to_append = np.array([19, 21, 18])

Step 4: Appending the array

With the help of the np.append() function, we can now append a row_to_append to empty_array.

row_appended_array = np.append(empty_array, row_to_append)

The final code looks like the below:

import numpy as np

empty_array = np.array([])

row_to_append = np.array([19, 21, 18])

row_appended_array = np.append(empty_array, row_to_append)

print(row_appended_array)

# Output: [19. 21. 18.]

If you are doing lots of appending, then the np.append() method is inefficient because it returns a new array each time you call.

If the empty_array and row_to_append have different data types, NumPy will upcast to the most general type to accommodate both.

Using Python list over Numpy array

If you are looking for efficient appending, consider Python lists, append elements there, and, in the end, convert it to a numpy array.

import numpy as np

data = []
for i in range(10):
    data.append([i, i*2])

final_array = np.array(data)
print(final_array)

Output

[[ 0  0]
 [ 1  2]
 [ 2  4]
 [ 3  6]
 [ 4  8]
 [ 5 10]
 [ 6 12]
 [ 7 14]
 [ 8 16]
 [ 9 18]]

If you are working with a multi-dimensional array, you can use the “axis” argument of the np.append() function.

This approach is highly recommended for appending elements frequently because it saves time.