Deepak, to be sure, I was referring to sequential guarantees with the longs.
I would suggest being careful with taking half the UUID as the probability of collision can be unexpectedly high. Many bits of the UUID is typically time-based so collision among those bits is virtually guaranteed with probability 1 when parallelized. Even if you can optimistically find some 64 uniformly random bits to use, due to the birthday paradox, the collision probability among 1 billion (2^32) values is something like 1 - exp(-1/2), or a very uncomfortable 40%. If you have orders of magnitude fewer edges/vertices, you'd have a wider margin of safety---but estimate it to be sure. -- Christopher T. Nguyen Co-founder & CEO, Adatao <http://adatao.com> linkedin.com/in/ctnguyen On Mon, Feb 24, 2014 at 12:38 PM, Deepak Nulu <deepakn...@gmail.com> wrote: > Thanks Christopher, I will look into the StackOverflow suggestion of > generating 64-bit UUIDs in the same fashion as 128-bit UUIDs. > > > > > -- > View this message in context: > http://apache-spark-user-list.1001560.n3.nabble.com/GraphX-with-UUID-vertex-IDs-instead-of-Long-tp1953p1990.html > Sent from the Apache Spark User List mailing list archive at Nabble.com. >
