Question

(b) Suppose you are designing a search aggregator, which for a given query fetches search results from two different search engines and presents an intersection of the two search results. Here is a simplified version of this problem: Given two sorted integer arrays of lengths m and n, return a new array with elements that are present in both input arrays. The input array may contain duplicates, but there should be no duplicates in the output array. For example, if the input arrays are [17, 23, 23, 35, 43, 47, 69, 78, 80, 84, 84, 86] and [23, 35, 50], the output array should be [23, 35]. (i) Describe a brute force solution. What is the worst-case tight big-O time complexity of this brute solution? (ii) Describe a solution that has worst-case tight-O (mlog n) time complexity. (iii) Describe a solution that has worst-case tight-O (m + n) time complexity (iv) Between subparts ii andii) above, is one solution always better than the other (in terms of runtime)? If so, which one and why? If not, describe the situation (in terms of m and n) when one is better tharn the other.

Answer 1

li) The e fosce methodthis case is, each eleme m awnay A, check cohetheut this element occus i atwray B By lineaurly dteavch the clements n 8 Thos i aonay tB nas sipe n. The checking for intetse ctío η of one element will take oln) Jme ond seavchinaal the m etements of A oill take o/mn) Time four ts occuance m Since enture oh aaysed, so for each elemevti check ashe the Jime an 00tCh in? B f0oy οηe, element aïll take 0(logn) Seaurchin take omlog n) Jime - ustlike m meoge olmtm) , we do the analoqous thmq Below is the code inte section (A B , 'nmi m) 十 -Se t -S: Empty and t-1hsets ill contain intewsection Yesult

omd A mot in S -- Then add A to S amd i++. j+t ktt 1 hen Kincrement index of 3 cise 10 As it con be seen This ato ile Ton in Time omi Sie of aasa The 30 and η iS Very @arge, andse -e m is very small Klue of o(-mlogn) will be less than o(mtn) and/Si3cơf omtogn) ill be less afe almost some the approach iv 1's better าทุ and