XML FeedsGoogle Interview Questions
So the economy is really bad and your job might be on the line, Google is always looking for new talent. Here are 2 questions that they asked a friend of mine. What would you answer?
1. You have been shrunk down to the size of a nickel and tossed into a blender. You are told that the blender blades will start in 60 seconds.What would you do to save your life?
2. Design and describe a system/application that will most efficiently produce a report of the top 1 million Google search requests.These are the particulars.
* You are given 12 servers to work with. They are all dual-processor machines with 4Gb of RAM, 4x400GB hard drives and networked together.(Basically, nothing more than high-end PC’s)
* The log data has already been cleaned for you. It consists of 100 Billion log lines, broken down into 12 320 GB files of 40-byte search terms per line.
* You can use only custom written applications or available free open-source software.
So what do you think, could you answer these questions?
And don’t forget, if you have any tech questions then don’t hesitate to ask in our forums
Denis has been working with SQL Server since version 6.5. Although he worked as an ASP/JSP/ColdFusion developer before the dot com bust, he has been working exclusively as a database developer/architect since 2002. In addition to English, Denis is also fluent in Croatian and Dutch, but he can curse in many other languages and dialects (just ask the SQL optimizer) He lives in Princeton, NJ with his wife and three kids.
Number 1 is pretty easy
Question 1:
Why would I save my life? I'm just a speckle on the face of this earth. 
relevancy is a virtue. 
question 1, start running around the inside of the blender. using centrifugal force, i would slowly work my way up and out of the top.
possiblities..for question 1
Can we get more questions on the same? 
Hi,
how silly .. I would google the answer. geez. 
1. Lay down, you can fit under the blades if you are a the size of a nickle. (hopefully)
Hi,
Great description of your interview process, I enjoyed reading it. 
Hi,
When you look into a fully loaded, running blender, it makes a whirlpool with a toroidal flow that descends in the center and ascends on the outer portions (nearer the walls of the blender vessel). 
1)I will turn myself into a nickel coin and sacrifice my tail. 
Nr. 2 is a textbook example of a MapReduce. So I would either use their internal one (had I gotten the job ;)) or Hadoop. Creating an application which could be run in a cluster of 12 CPUs would be a breeze and the rest would be simply the logic behind report generation.
I don't understand, how will answering these questions reveal to Google that you're a good programmer? The first question is obviously written by someone who has some serious issues... 
#2 One solution is use unix sort and uniq. So on each server you put an 26 GB file. Sorting in memory is the fastest so you use unix split to make the files somewhat less than 4GB. Then sort each file. Concat the files and do sort so there is one file left. That will be the slowest part. Now you can do a unix uniq to get the counts. Then you move the top 1M results with the counts from each server to one server. Sort by search terms, sum counts per unique search term, sort by count. Do a head -1000000 on the file.
Good questions. 
1) I'll stay in one of these places at the bottom where I always find raw stuff left, when the blending is over...
LTD Social Sitings
Note: Watch for social icons on posts by your favorite authors to follow their postings on these and other social sites.