Filter Before You Parse: Faster Analytics on Raw Data with Sparser

Sounds a lot like the &quot;on the fly parsing&quot; (§3.1) in Alagiannis&#x27; NoDB (See <a href="https:&#x2F;&#x2F;stratos.seas.harvard.edu&#x2F;files&#x2F;stratos&#x2F;files&#x2F;nodb-cacm.pdf" rel="nofollow">https:&#x2F;&#x2F;stratos.seas.harvard.edu&#x2F;files&#x2F;stratos&#x2F;files&#x2F;nodb-ca...</a> for details).

I’m more interested in the work some folks are doing on using succinct structures style techniques to accelerate parsing. <a href="https:&#x2F;&#x2F;github.com&#x2F;haskell-works&#x2F;hw-json" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;haskell-works&#x2F;hw-json</a><p>It’s still relatively immature. But it’s a more algorithmic approach that I think plays nice with pretty much any source of semistructred Data. Though simd acceleration certainly is pretty sweet too

Anyone figure out how to get an instance of spark up with sparser working?

#define PREPROCESSING 1

I have done that for a long time.