I concurrently applied for an opening on Google's YouTube team and was put in contact with a recruiter by an old neighbor.

The YouTube application turned into a 45-minute phone interview, which turned into four onsite interviews of similar length.

The interviews focused on problem-solving and technical questions (as opposed to free-form questions like "What is your biggest strength/weakness?").

The setup is that we are given a series of text files which contain information regarding a code repository's commits. Each file represents a single commit and they are formatted as follows:

Commit #: XXX Author: XXX Reviewer(s): XXX, XXX, ... File: XXX File: XXX ... Date: XX:XX XX/XX/XXXX

The commit number is unique and is generated in synchronous order. There is exactly 1 unique author. There are a variable number of reviewers, delimited by commas; if there are no reviewers, that line is absent from the file. There are a variable number of edited files in the commit, each receiving its own line. The time/date is when the commit was submitted.

First, design a graphical model for all of the commit data. Then, describe how this model is updated when a new commit is generated. Finally, write the code segment called when a new commit is generated which edits a system that has implemented your model of the data. Its input is a file name and whatever necessary data structures are maintained by your system.