How do I practice coding and algorithm questions?

Use PracHub's coding console to write, test, and debug your solutions in Python or JavaScript. View hints, test against sample inputs, and compare with official solutions.

What difficulty level is this coding question?

This is a medium difficulty Coding & Algorithms question, commonly asked during Onsite rounds at Databricks.

What role is this question designed for?

This question is commonly asked for Machine Learning Engineer candidates at Databricks during technical interviews.

Implement a Lazy Array | Databricks Coding Question

Quick Overview

This question evaluates understanding of lazy evaluation and deferred computation, implementation of iterator/generator-like abstractions, and the ability to compose functional transformations (map/filter) while managing performance and memory trade-offs.

Implement a Lazy Array

Company: Databricks

Role: Machine Learning Engineer

Category: Coding & Algorithms

Difficulty: medium

Interview Round: Onsite

Implement a lazily evaluated array abstraction. The object should wrap an underlying sequence and support chained transformations such as `map` and `filter` without executing them immediately. Computation should only happen when the result is materialized, for example through iteration or a method like `to_list()`. After implementing it, explain how you would test that the implementation is truly lazy rather than eagerly computing intermediate results.

Quick Answer: This question evaluates understanding of lazy evaluation and deferred computation, implementation of iterator/generator-like abstractions, and the ability to compose functional transformations (map/filter) while managing performance and memory trade-offs.

Implement a lazily evaluated array abstraction over a sequence of integers. The abstraction should support chained transformations such as map and filter without creating intermediate arrays immediately. Computation must only happen when the result is materialized, such as through iteration or a method like `to_list()`. For this problem, write `solution(nums, operations)` that builds a lazy pipeline over `nums`, applies the operations from left to right, and returns the final materialized list. Supported operations: - `('map', 'add', x)`: replace each value `v` with `v + x` - `('map', 'mul', x)`: replace each value `v` with `v * x` - `('map', 'sub', x)`: replace each value `v` with `v - x` - `('filter', 'gt', x)`: keep values greater than `x` - `('filter', 'lt', x)`: keep values less than `x` - `('filter', 'ge', x)`: keep values greater than or equal to `x` - `('filter', 'le', x)`: keep values less than or equal to `x` - `('filter', 'eq', x)`: keep values equal to `x` - `('filter', 'even')`: keep even values - `('filter', 'odd')`: keep odd values - `('take', k)`: keep only the first `k` values that reach this point in the pipeline A correct implementation should store transformations and only execute them when the final result is requested. How to test true laziness: use map/filter functions with side effects, such as incrementing a counter or raising an exception when an element is touched. Before iteration or `to_list()`, the counter should remain 0. Also test a pipeline ending with `take(k)` and verify that only the minimum number of source elements needed to produce `k` results are evaluated.

Constraints

0 <= len(nums) <= 100000
0 <= len(operations) <= 10000
-10^9 <= nums[i], x <= 10^9
For `('take', k)`, 0 <= k <= 100000
All operations are valid and appear in one of the supported forms

Loading coding console...

Quick Overview

Examples

Hints