np.choose() for conditional selection in NumPy

Introduction

We use np.choose() to pick values from multiple options based on conditions. It helps select data easily without writing many if-else statements.

You want to assign categories to numbers based on ranges.

You need to select different arrays depending on a condition array.

You want to replace values in an array based on multiple conditions.

You want a simple way to map indices to values in data analysis.

Syntax

NumPy

np.choose(index_array, [option1, option2, option3, ...])

index_array contains integers that act as indices to pick from the options list.

All options must have the same shape or be broadcastable to the same shape.

Examples

This picks values from three arrays based on index. For example, index 0 picks from choices, index 1 picks from choices*2, etc.

NumPy

import numpy as np

choices = np.array([10, 20, 30])
index = np.array([0, 1, 2])
result = np.choose(index, [choices, choices*2, choices*3])
print(result)

Here, cond selects which array value to pick for each position.

NumPy

import numpy as np

cond = np.array([1, 0, 2, 1])
options = [np.array([100, 100, 100, 100]), np.array([200, 200, 200, 200]), np.array([300, 300, 300, 300])]
result = np.choose(cond, options)
print(result)

Sample Program

This program assigns grades to scores using np.choose(). We first create an index array where each score is mapped to 0, 1, or 2 based on its value. Then, np.choose() picks the grade label for each score.

NumPy

import numpy as np

# Suppose we have scores and want to assign grades:
# 0 = Fail, 1 = Pass, 2 = Excellent
scores = np.array([55, 70, 85, 40, 95])

# Create an index array based on score ranges
index = np.where(scores < 60, 0, np.where(scores < 80, 1, 2))

# Define grade labels
grades = np.array(['Fail', 'Pass', 'Excellent'])

# Use np.choose to pick grade labels
result = np.choose(index, [grades[0], grades[1], grades[2]])

print(result)

Important Notes

The index_array must contain integers starting from 0 up to the number of options minus one.

If the shapes of options differ, NumPy tries to broadcast them to a common shape.

Use np.where() to create the index array for complex conditions before using np.choose().

Summary

np.choose() selects elements from multiple arrays based on an index array.

It simplifies conditional selection without many if-else statements.

Make sure the index array and options have compatible shapes.

Practice

1. What does the np.choose() function do in NumPy?

easy

A. Sorts an array in ascending order

B. Selects elements from multiple arrays based on an index array

C. Calculates the mean of an array

D. Generates random numbers

Solution

1. Step 1: Understand the purpose of np.choose()

   np.choose() uses an index array to pick elements from multiple arrays at corresponding positions.

2. Step 2: Compare with other options

   Sorting, mean calculation, and random number generation are unrelated to np.choose().

3. Final Answer:

   Selects elements from multiple arrays based on an index array -> Option B

4. Quick Check:

   np.choose() = Select elements [OK]

Hint: Remember: np.choose picks from arrays using index array [OK]

Common Mistakes:

• Confusing np.choose() with sorting functions
• Thinking np.choose() calculates statistics
• Assuming np.choose() generates random values

2. Which of the following is the correct syntax to use np.choose() with an index array idx and option arrays a, b, c?

easy

A. np.choose(idx, [a, b, c])

B. np.choose([a, b, c], idx)

C. np.choose(a, b, c, idx)

D. np.choose(idx, a, b, c)

Solution

1. Step 1: Recall np.choose() parameter order

   The first argument is the index array, the second is a list of arrays to choose from.

2. Step 2: Match syntax with options

   Only np.choose(idx, [a, b, c]) matches the correct order and structure.

3. Final Answer:

   np.choose(idx, [a, b, c]) -> Option A

4. Quick Check:

   Index first, then list of arrays [OK]

Hint: Index array first, then list of arrays in np.choose() [OK]

Common Mistakes:

• Swapping index array and options
• Passing options as separate arguments instead of a list
• Using incorrect argument order

3. What is the output of the following code?

import numpy as np
idx = np.array([0, 2, 1, 0])
a = np.array([10, 20, 30, 40])
b = np.array([100, 200, 300, 400])
c = np.array([1000, 2000, 3000, 4000])
result = np.choose(idx, [a, b, c])
print(result)

medium

A. [10 2000 300 40]

B. [10 300 200 40]

C. [100 3000 2000 400]

D. [1000 300 200 40]

Solution

1. Step 1: Understand index array and options

   Index array is [0, 2, 1, 0]. Arrays are a=[10,20,30,40], b=[100,200,300,400], c=[1000,2000,3000,4000].

2. Step 2: Select elements using np.choose()

   For each position, pick from arrays based on index:
   pos0: idx=0 -> a[0]=10
   pos1: idx=2 -> c[1]=2000
   pos2: idx=1 -> b[2]=300
   pos3: idx=0 -> a[3]=40
   So result = [a[0], c[1], b[2], a[3]] = [10, 2000, 300, 40]

3. Step 3: Check options

   [10 300 200 40] incorrectly uses 300 and 200 instead of 2000 and 300. [10 2000 300 40] matches our result.

4. Final Answer:

   [10 2000 300 40] -> Option A

5. Quick Check:

   np.choose picks arrays by index, then elements by position [OK]

Hint: Index picks array, position picks element [OK]

Common Mistakes:

• Mixing up index meaning (array vs element)
• Swapping elements from wrong arrays
• Confusing np.choose with fancy indexing

4. The following code raises an error. What is the most likely cause?

import numpy as np
idx = np.array([0, 1, 2])
a = np.array([1, 2])
b = np.array([3, 4])
c = np.array([5, 6])
result = np.choose(idx, [a, b, c])
print(result)

medium

A. np.choose() requires index array to be boolean

B. Index array contains values outside the range of options

C. Index array and option arrays have incompatible shapes causing mismatch

D. np.choose() cannot handle more than two option arrays

Solution

1. Step 1: Check index array and options

   Index array is [0,1,2], which is valid for 3 options (a,b,c).

2. Step 2: Check shapes of option arrays

   a, b, c each have shape (2,), but index array has length 3. np.choose expects index and options to have compatible shapes.

3. Step 3: Identify cause of error

   Index array length 3 vs option arrays length 2 causes shape mismatch error.

4. Final Answer:

   Index array and option arrays have incompatible shapes causing mismatch -> Option C

5. Quick Check:

   Shape mismatch causes np.choose error [OK]

Hint: Ensure index and option arrays have compatible shapes [OK]

Common Mistakes:

• Assuming index values cause error
• Thinking np.choose needs boolean index
• Believing np.choose limits number of options

5. You have three arrays representing scores from three different tests:

test1 = np.array([80, 90, 70])
test2 = np.array([85, 88, 75])
test3 = np.array([78, 92, 72])

And an index array indicating the best test for each student:

best = np.array([1, 2, 0])

Using np.choose(), how do you create an array of the best scores for each student?

hard

A. np.choose(best, [test1, test3, test2])

B. np.choose(best, [test3, test1, test2])

C. np.choose(best, [test2, test3, test1])

D. np.choose(best, [test1, test2, test3])

Solution

1. Step 1: Understand index array meaning

   Index array best = [1, 2, 0] means:
   For student 0, best test is index 1 -> test2
   For student 1, best test is index 2 -> test3
   For student 2, best test is index 0 -> test1

2. Step 2: Match option arrays order to index

   To pick correctly, option arrays must be in order [test1, test2, test3] so index 0 picks test1, index 1 picks test2, index 2 picks test3.

3. Step 3: Verify np.choose usage

   Using np.choose(best, [test1, test2, test3]) picks correct scores per student.

4. Final Answer:

   np.choose(best, [test1, test2, test3]) -> Option D

5. Quick Check:

   Order options to match index meaning [OK]

Hint: Order option arrays to match index values [OK]

Common Mistakes:

• Assuming option arrays order matches test numbers
• Not matching index values to option array order
• Using wrong order causing incorrect selection