sf-wdi-37 / whiteboarding

[whiteboarding]

Whiteboarding Interview Practice

Process

Pair up! One person will act as the interviewer, while the other will act as the interviewee. Switch off after each question. The interviewer may reference the solutions and give hints.

Be thinking about Cracking the Coding Interview's "Five Steps to a Technical Question", paraphrased below:

1. Ask your interviewer questions to resolve any ambiguities.
2. Design an Algorithm.
3. Write pseudocode first (but tell your interviewer you plan to write real code later).
4. Write your code at a moderate pace.
5. Test your code with test cases and carefully fix any mistakes.

Also, talk out loud! Your interviewer is trying to learn how you think through problems.

Warmup

1. Given 2 arrays of the same length, create a Hash (in Ruby) or Object (in JavaScript) where the elements from the first array are used as keys and elements from the second array are used as values.

   to_hash [:name, :age, :height], ["bill", 47, 70]
   # => {:name => "bill", :age => 47, :height => 70}

What is the runtime (big o) of your solution?

2. Given a sorted array and a particular value, write a function to find the index of that value in the array. Don't use arr.indexOf(val) (in JavaScript) or arr.index(val) (in Ruby). If the value is not in the array, return -1.

search([1,4,7,9,14,15], 7)
// 2
  
search([1,4,7,9,14,15], 199)
// -1

What is the runtime (big o) of your solution?

Practice

1. Write a function to 'compress' a string by counting repeated characters. For example, the string "ooohmmmmmmm" would become "o3h1m7". If your compressed string would be longer than the original, you should return the original instead.

What is the runtime (big o) of your solution?

2. Given two strings, check if the first string occurs within the second string. Return true if it does and false if not.

Hint: you can use built-in methods.

substring?("abc", "catabcd")
# true
substring?("a", "eeee")
# false

What is the runtime (big o) of your solution?

What if you don't use built-in methods?

Fast algorithms for this problem run in O(n + k) time or even O(n) time, where n is the length of the longer string and k is the length of the shorter. In the tradeoff, they use extra storage space. This is the "(sub)string search" problem.

Advanced

1. Matrix Search: Given an array of arrays (called a matrix) that is sorted from the top left to the bottom right, find a given value within the matrix. Return a pair of numbers that indicate the value's row and column. If the value is not present, return [-1, -1]. Can you use the idea of binary search to do it quickly?

matrix = [
	[  1,  3,  6,  9 ],
	[ 10, 14, 18, 20 ],
	[ 28, 30, 34, 37 ]
]
searchNested(matrix, 18)
// [1,2]
searchNested(matrix, 17)
// [-1,-1]

What is the runtime (big o) of your solution?

A more difficult variation of this problem is described and explained by someone named 1337c0d3r, here.

2. Given a list of tasks with prerequisites, return an order in which the tasks can be completed.

task_list = [
  { id: 1, description: "put on pants", prerequisites: [4] },
  { id: 2, description: "put on shirt", prerequisites: [4] },
  { id: 3, description: "put on shoes", prerequisites: [1, 5] },
  { id: 4, description: "put on undergarmets", prerequisites: [] },
  { id: 5, description: "put on socks", prerequisites: [] }
  
]
order(task_list)
# there are many valid orders including:
# => [4, 5, 1, 2, 3]

What is the runtime (big o) of your solution?

This problem is "topological sort," if you'd like to look it up. It's used in scheduling and for things like compiling all of a package's dependencies.

3. Given two strings, find their longest common substring, the longest series of characters that they share.

su = "supercalifragilisticespialidocious"
  ca = "californian"
  xy = "xyzzy"
longestCommonSubstring(su, ca)
// "calif"
longestCommonSubstring(ca, xy)
// ""

What is the runtime (big o) of your solution?

This is also known as the "longest common subsequence" problem, if you'd like to look it up.