[[1, 2, 3], [4, 5, 6]] and you want to convert this into a vector of form [1, 2, 3, 4, 5, 6]. This article explores effective methods to achieve this transformation.Method 1: Using NumPy’s Flatten Method
Flattening is the process of converting a multi-dimensional array into a one-dimensional array. NumPy provides the flatten() method to achieve this. When the flatten method is applied to an array, it collapses the array across all dimensions and returns a new one-dimensional array with all the original elements.
Here’s an example:
♥️ Info: Are you AI curious but you still have to create real impactful projects? Join our official AI builder club on Skool (only $5): SHIP! - One Project Per Month
import numpy as np # Create a 2D numpy array array_2d = np.array([[1, 2, 3], [4, 5, 6]]) # Flatten the array vector = array_2d.flatten() print(vector)
Output:
[1 2 3 4 5 6]
In the above code snippet, the method flatten() was called on a 2D NumPy array, which returned a new 1D array (a vector) containing all the elements from the original array in the same order.
Method 2: Using NumPy’s Ravel Method
NumPy’s ravel() method is similar to flatten(), but it returns a flattened one-dimensional array that is a view of the input array whenever possible. This means that the returned array often shares the same memory space as the input array which can make this method more memory efficient.
Here’s an example:
import numpy as np # Create a 2D numpy array array_2d = np.array([[7, 8, 9], [10, 11, 12]]) # Use ravel to create a vector vector = array_2d.ravel() print(vector)
Output:
[7 8 9 10 11 12]
Using ravel(), we got a 1D array with shared memory, which is efficient. However, since it’s a view, modifying the returned array may also modify the original array.
Method 3: Using NumPy’s Reshape Method
The reshape() function in NumPy is a versatile tool for transforming the shape of an array. It can be used to flatten an array by specifying the new shape as (-1,), which tells NumPy to infer the correct dimension for a one-dimensional vector.
Here’s an example:
import numpy as np # Create a 2D numpy array array_2d = np.array([[13, 14, 15], [16, 17, 18]]) # Reshape the array into a vector vector = array_2d.reshape(-1) print(vector)
Output:
[13 14 15 16 17 18]
The reshape() method is used to convert the 2D array into a 1D array without changing the data. The argument -1 tells NumPy to calculate the appropriate length for the new dimension.
Method 4: Using NumPy’s Squeeze Method
The squeeze() method is particularly useful when you have an array with one or more dimensions of size one. This method removes such singleton dimensions and returns an array with fewer dimensions. If you start with a 2D array with one of the dimensions being 1, squeeze() will return a vector.
Here’s an example:
import numpy as np # Create a 2D numpy array with one singleton dimension array_2d = np.array([[19, 20, 21]]) # Squeeze the array to remove singleton dimensions vector = array_2d.squeeze() print(vector)
Output:
[19 20 21]
Here, squeeze() removed the singleton dimension, resulting in a vector. It’s a simple and clean way to remove unnecessary dimensions in an array.
Bonus One-Liner Method 5: Using NumPy’s Flat Iterator
NumPy arrays have an attribute called flat, which is an iterator over all the items in the array. While not conventionally used to convert arrays to vectors, it can be used in a list comprehension to create a vector.
Here’s an example:
import numpy as np # Create a 2D numpy array array_2d = np.array([[22, 23, 24], [25, 26, 27]]) # Use flat attribute to create a vector vector = np.array([item for item in array_2d.flat]) print(vector)
Output:
[22 23 24 25 26 27]
We iterate over each item using the flat attribute and create a list, which we then convert to a NumPy array, getting the 1D vector as expected.
Summary/Discussion
- Method 1: Flatten. Creates a new, contiguous one-dimensional array. Simple and reliable. May require additional memory for the copy.
- Method 2: Ravel. It usually returns a view of the original array, so it is memory-efficient. Modifications can affect the original array, which may be undesirable.
- Method 3: Reshape. Allows for flexible dimension changes. It can sometimes return a view or a copy, so behavior with respect to memory can vary.
- Method 4: Squeeze. Good for removing singleton dimensions. Not strictly for flattening, but can be useful in specific cases where an array has unnecessary dimensions of unit size.
- Bonus Method 5: Flat Iterator. Can be powerfully combined with list comprehensions. It is more verbose but offers fine-grained control over array transformation.