dplyr and tessera

[Enzo]

Hi

I am trying to explain to colleagues the difference between dplyr and tessera.

Taking into account the effort to produce a version of dplyr compatible with Spark, it would appear that the main differences are:

- hadoop / mapreduce (dplyr does not offer mapreduce out-of-the-box) 

- trelliscope

Did I get this right?  Is there a better / more precise way to explain this?

[Ryan Hafen]

Hi Enzo,

There is a section dedicated to this here: http://tessera.io/docs-datadr/#faq.  See: "How is datadr similar to / different from plyr / dplyr?"

Basically it comes down to scale and flexibility.  You can scale to much larger data sets and have much more flexibility in data structures and algorithms you apply in datadr.  This has to do with the fact that you are using a distributed key-value store and mapreduce for storage and computation.  When data is large and complex, it is often not too well suited to be stored in a tidy tabular format, and flexibility over data structures is much-needed.  I should mention that in the distributed data frame (ddf) case, I do have plans to add a dplyr interface to datadr.

Also, you are correct that trelliscope is a differentiator, although we have added a utility function to be able to use a grouped dplyr table as input to trelliscope, as illustrated in this blog post: http://tessera.io/blog/quick-interactive-faceting/

Ryan

--
You received this message because you are subscribed to the Google Groups "Tessera-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tessera-user...@googlegroups.com.
To post to this group, send email to tesser...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tessera-users/6944e107-5236-43a4-9cc6-9be830015e27%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.