MariaDB Corp. has announced that release 2.0 of its MaxScale database proxy software is henceforth no longer open source. The organization has made it source-available under a proprietary license that promises each release will eventually become open source once it’s out of date.
MaxScale is at the pinnacle of MariaDB Corp.’s monetization strategy — it’s the key to deploying MariaDB databases at scale. The thinking seems to be that making it mandatory to pay for a license will extract top dollar from deep-pocketed corporations that might otherwise try to use it free of charge. This seems odd for a company built on MariaDB, which was originally created to liberate MySQL from the clutches of Oracle.
According to the results of the 2016 survey, R is the preferred tool for 42% of analytics professionals, followed by SAS at 39% and Python at 20%. While Python’s placing may at first appear to relegate the language to Bronze Medal status, it’s the delta here that really matters.
It’s interesting to see the breakdowns of who uses which language, comparing across industry, education, work experience, and geographic lines.
If your data source is a SQL database, you might have implemented pagination by using
LIMIT .. OFFSET, or
OFFSET .. FETCHor some
ROWNUM / ROW_NUMBER()filtering (see the jOOQ manual for some syntax comparisons across RDBMS). OFFSET is the right tool to jump to page 317, but remember, no one really wants to jump to that page, and besides, OFFSET just skips a fixed number of rows. If there are new rows in the system between the time page number 316 is displayed to a user and when the user skips to page number 317, the rows will shift, because the offsets will shift. No one wants that either, when they click on “next”.
Instead, you should be using what we refer to as “keyset pagination” (as opposed to “offset pagination”).
He also has a good explanation of the seek method.
I will throw in one jab at Oracle (because hey, it’s been a while since I’ve lobbed a bomb at Oracle on this blog): it’d really suck to have a system where I legally wasn’t allowed to distribute relevant performance comparison benchmarks. Fortunately, I tend to work on better data stacks.
Why is this code better than the functional approach? Note that it saves an enormous amount of time since this approach does not need to scan through all the integers! It are simply a few calculations (at which a computer is good at). All the code (the naive approach and the better approach) can be found on GitHub.
Having a solid understanding of mathematics and logic can help you come up with superior algorithms, but make sure you comment them in detail so that the next dev (who might not understand the underpinnings of your code) doesn’t replace it with a brute-force method because it’s “easier.”