Baidu Interview Question: Given a 10G user action log f... | Glassdoor

Interview Question

Software Engineer Interview Beijing, Beijing (China)

Given a 10G user action log file, and each each in the log

  is userID-actionID format, also give you a PC with 4G memory, find a way to search for those userIDs that has at least three action records.
Answer

Interview Answer

1 Answer

0

I failed in this question, I cannot find a way to do it beside use external sort to sort the file, and go through the file to find the record. However, I think this can be solve by using map reduce idea on a single machine(or hadoop on single node). (M/R and hadoop can be found on wiki)

Interview Candidate on Oct 3, 2011

Add Answers or Comments

To comment on this, Sign In or Sign Up.