Re: How to handle Big Data under IBM i?

From: Matt Olson
I don't need to regurgitate already available documentation,
please use google for more insight.

From only your source code, it wasn't THAT clear what to search on. I finally came upon a Microsoft reference to SqlBulkCopy, which appears to be quite comparable to the IBM i CPYF command. Here's a CL program that appends 1,000,000 rows from one table to another:

http://pastebin.com/5c0snQN8


Here's an RPG program that calls the CL from within a loop 130 times (to generate 130,000,000 rows):

http://pastebin.com/tKeJ6r15


That entire process completes in 93 seconds, so there's evidently some optimization to be achieved by doing bulk copies of physical file members.

But these kinds of exercises are something of a distraction from the topic of handling Big Data because they are simply tests of single threaded jobs running against single database tables. Products like MemSQL and VoltDB are mostly about partitioning tables into smaller chunks and updating and reading them from parallel servers.

-Nathan