On Wed, Dec 9, 2009 at 7:48 PM, Aditya Shankar <iitm.adityashan...@gmail.com > wrote:
> Hi, > > On Wed, Dec 9, 2009 at 2:53 PM, Manisha <pgo...@gmail.com> wrote: > >> We have a server hit by millions of users. Sever log files contains >> the user ids of all of them. How do we find the frequency of login of >> each user. What will the most efficient way to store the users, and >> access them to find their frequency(The log files are very huge!!) >> > oops... > Can't we just use a hash table? It's O(1) access, addition and updation > (and removal for that matter). > There will not be conflicts because the user IDs are presumably unique. > >> >> I thought of using B+ tree indexing with user ids as the key. Leaf >> nodes will have the pointers to bucket of user ids. One item of bucket >> will contain user id and frequency of this user. >> For insertion, search complexity will be ~O(logn) >> >> Any potential problem with approach? Are there any better approach to >> tackle this problem? >> >> -- >> >> You received this message because you are subscribed to the Google Groups >> "Algorithm Geeks" group. >> To post to this group, send email to algoge...@googlegroups.com. >> To unsubscribe from this group, send email to >> algogeeks+unsubscr...@googlegroups.com<algogeeks%2bunsubscr...@googlegroups.com> >> . >> For more options, visit this group at >> http://groups.google.com/group/algogeeks?hl=en. >> >> >> > -- You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To post to this group, send email to algoge...@googlegroups.com. To unsubscribe from this group, send email to algogeeks+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/algogeeks?hl=en.
