29 Mar Joining Numpy Arrays

We can join two or more arrays in a single new array using the following:

1. concatenate() method: Joining along with axis, unlike keys in SQL.
2. stack method: Joining along a new axis, unlike keys in SQL.

In NumPy, axis=0 refers to the index (rows), and axis=1 refers to columns. Works differently in case of concatenate(), stack(), and aggregation (sum, mean, etc.) methods. The behavior depends on whether you are joining data (concatenating/stacking) or summarizing data (aggregating).

Let us see the comparison:

Joining Numpy Arrays with concatenate()

In NumPy, concatenate() is used to join two or more arrays along an existing axis. It does not create a new axis but merges arrays as per the specified dimension.

Syntax

Here,

• (array1, array2, …) → A sequence (tuple/list) of arrays you want to join.
• axis → The dimension along which arrays are joined. Default is 0 (rows).

Return: A new array formed by joining the input arrays along the specified axis.

Let us understand the axis parameter of the concatenate() method:

axis=0 → Vertical stacking (row-wise)

• Concatenates along the first axis (rows).
• Arrays are stacked on top of each other.
• Requires that the arrays have the same number of columns.

axis=1 → Horizontal stacking (column-wise)

• Concatenates along the second axis (columns).
• Arrays are placed side by side.
• Requires that the arrays have the same number of rows

Let us now see two examples:

Example 1

# Join NumPy Arrays with concatenate()
# Example 1

import numpy as np

# Create two arrays
n1 = np.array([[5, 10, 15, 20]])
n2 = np.array([[25, 30, 35, 50]])

print(n1)
print(n1.ndim)
print(n1.shape)

print(n2)
print(n2.ndim)
print(n2.shape)

# Stack vertically (row-wise)
# Concatenates along the first axis(rows)
res = np.concatenate((n1, n2), axis = 0)
print(res)
print(res.ndim)
print(res.shape)

# Stack horizontally (column-wise)
# Concatenates along the second axis(columns)
res = np.concatenate((n1, n2), axis = 1)
print(res)
print(res.ndim)
print(res.shape)

Output

[[ 5 10 15 20]]
2
(1, 4)
[[25 30 35 50]]
2
(1, 4)
[[ 5 10 15 20]
 [25 30 35 50]]
2
(2, 4)
[[ 5 10 15 20 25 30 35 50]]
2
(1, 8)

Example 2

# Join NumPy Arrays with concatenate()
# Example 2

import numpy as np

n1 = np.array([[1, 2],
               [3,4]])
n2 = np.array([[5, 6],
               [7,8]])

print(n1)
print(n1.ndim)
print(n1.shape)

print(n2)
print(n2.ndim)
print(n2.shape)

print("----------------------")

res = np.concatenate((n1, n2), axis=0)
print(res)
print(res.ndim)
print(res.shape)

res = np.concatenate((n1, n2), axis=1)
print(res)
print(res.ndim)
print(res.shape)

Output

[[1 2]
 [3 4]]
2
(2, 2)
[[5 6]
 [7 8]]
2
(2, 2)
----------------------
[[1 2]
 [3 4]
 [5 6]
 [7 8]]
2
(4, 2)
[[1 2 5 6]
 [3 4 7 8]]
2
(2, 4)

Joining Numpy Arrays with Stack Methods

Join arrays in Numpy using the stack method, such as

• stack(): Combines arrays along a new axis (adds dimension).
• hstack(): Stacks arrays horizontally (column-wise, side by side).
• vstack(): Stacks arrays vertically (row-wise, one below another).
• dstack(): Stacks arrays depth-wise (along third axis).
• column_stack(): Stacks 1D arrays as columns into a 2D array.

Each one differs by axis of stacking: adds a new axis, while hstack / vstack/ dstack/ column_stack arrange along existing ones

Let us see the examples one by one:

numpy.stack() in Python

Join Numpy arrays using the stack() method. It combines arrays along a new axis (adds dimension).

# Join NumPy arrays using stack() method

import numpy as np

n1 = np.array([1, 2, 3])
n2 = np.array([4, 5, 6])
print(n1.ndim)
print(n2.ndim)

# Join
# Here, axis = 0 is the default
res = np.stack((n1, n2), axis = 0)
print(res)
print(res.ndim)

# Join
res = np.stack((n1, n2), axis = 1)
print(res)

Output

1
1
[[1 2 3]
 [4 5 6]]
2
[[1 4]
 [2 5]
 [3 6]]

numpy.hstack() in Python

Stack along rows using the hstack() method in Numpy:

# Join NumPy arrays using hstack() method

import numpy as np

n1 = np.array([1, 2, 3])
n2 = np.array([4, 5, 6])
print(n1.ndim)
print(n2.ndim)

# stack along rows
res = np.hstack((n1, n2))
print(res)
print(res.ndim)

Output

1
1
[1 2 3 4 5 6]
1

numpy.vstack() in Python

Stack along columns using the vstack() method in Numpy:

# Join NumPy arrays using vstack() method

import numpy as np

n1 = np.array([1, 2, 3])
n2 = np.array([4, 5, 6])
print(n1.ndim)
print(n2.ndim)

# stack along columns
res = np.vstack((n1, n2))
print(res)
print(res.ndim)

Output

1
1
[[1 2 3]
 [4 5 6]]
2

numpy.dstack() in Python

Stack along depth, i.e., height, using the dstack() method in Numpy. The dstack() stacks arrays in sequence depth-wise (along the third axis):

• It essentially adds a new third dimension and places arrays along it.
• It stacks arrays along axis = 2 (the third dimension).
• All input arrays must have the same shape along all but the third axis.
• If you pass 1D arrays, they are first reshaped to 2D before stacking.

Let us see two examples:

Example 1

# Join NumPy arrays using dstack() method
# Example 1

import numpy as np

n1 = np.array([1, 2, 3])
n2 = np.array([4, 5, 6])
print(n1)
print(n1.ndim)
print(n1.shape)
print(n2)
print(n2.ndim)
print(n2.shape)

# stack along depth (height)
# stacks arrays along the third dimension
res = np.dstack((n1, n2))
print(res)
print(res.ndim)
print(res.shape)

Output

[1 2 3]
1
(3,)
[4 5 6]
1
(3,)
[[[1 4]
  [2 5]
  [3 6]]]
3
(1, 3, 2)

Example 2

# Join NumPy arrays using dstack() method
# Example 2

import numpy as np

n1 = np.array([[1, 2],
               [3, 4]])
n2 = np.array([[5, 6],
               [7, 8]])
print(n1)
print(n1.ndim)
print(n1.shape)
print(n2)
print(n2.ndim)
print(n2.shape)

# stack along depth (height)
# stacks arrays along the third dimension
res = np.dstack((n1, n2))
print(res)
print(res.ndim)
print(res.shape)

Output

[[1 2]
 [3 4]]
2
(2, 2)
[[5 6]
 [7 8]]
2
(2, 2)
[[[1 5]
  [2 6]]

 [[3 7]
  [4 8]]]
3
(2, 2, 2)

numpy.column_stack() in Python

Stack according to columns using the column_stack() method in Numpy:

# Join NumPy arrays using column_stack() method

import numpy as np
n1 = np.array([1, 2, 3])
n2 = np.array([4, 5, 6])

print(n1)
print(n1.ndim)
print(n1.shape)
print(n2)
print(n2.ndim)
print(n2.shape)

# Join arrays
res = np.column_stack((n1, n2))
print(res)
print(res.ndim)
print(res.shape)

Output

[1 2 3]
1
(3,)
[4 5 6]
1
(3,)
[[1 4]
 [2 5]
 [3 6]]
2
(3, 2)

If you liked the tutorial, spread the word and share the link and our website Studyopedia with others.

For Videos, Join Our YouTube Channel: Join Now

---

Read More:

• Functions in Python with Examples
• Python Tuples Tutorial with Examples
• Dictionary in Python
• Python Lists
• DateTime in Python
• Python Numbers
• Type Conversion in Python