Segment data by client name and version

[will123195]

There could be different clients/versions pushing data to megafunnel. For example, a new version of condor could push extra csv columns or an iOS app could log custom events different than the condor csv.

So, megafunnel should require the leading csv columns to indicate the client name and version. This way we can change the condor csv schema at a later date.

Here's an example of data in megafunnel received from condor:

1411430134922,click,1280,855,0,4107,http://example.com/articles/95083/awefawef-awefawefawef-waefawefawef-nbsp,78802,"Mon, 22 Sep 2014 23:55:32 GMT",240,http://example.com/articles/94966/this-cat-is-unquestionably-frumpy,html.js.draganddrop.websockets.csstransitions.filereader > body.article.heroimage.logged-out,1207,4541,,,,,,0

I'm thinking instead it should have leading metadata columns:

1411430134922,condor,1.1.0,click,1280,855,0,4107,http://example.com/articles/95083/awefawef-awefawefawef-waefawefawef-nbsp,78802,"Mon, 22 Sep 2014 23:55:32 GMT",240,http://example.com/articles/94966/this-cat-is-unquestionably-frumpy,html.js.draganddrop.websockets.csstransitions.filereader > body.article.heroimage.logged-out,1207,4541,,,,,,0
              ^      ^

And then we can log all sorts of arbitrary stuff from different sources and it can be easily separated:

1411430134922,ios-app,0.2.0,login,1bd00ca3-1ab5-4865-aa0f-a6eb90a6cf43

@kesla @dominictarr @hij1nx thoughts?

[ralphtheninja]

👍

[dominictarr]

yes this is a good idea.

[kesla]

👍

[ralphtheninja]

This is going to be so bad ass it's unreal 👯