`df %>% assertr::assert(assertr::not_na, var)` gives incorrect message about assertr

[tonyfischetti]

Haha, so I hope this isn't too late but I fixed the namespace issue. And I'm writing the is_id (I think I'm going to use is_unique though) right now :)

[jjchern]

Perhaps assertr::assert_not_na(df, var) is too long.

How about assertr::nona(df, var)?

Also, Stata has a command isid that verifies if a variable, or a combination of several variables, is uniquely identified. Maybe assertr can have a similar function? Something like

assertr::isid(df, var)
assertr::isid(df, var1, var2)

I'm happy to submit a PR if you think these are useful functions to have.

Of course, nona and isid are not verbs, and they can be easily generated with assertr and magrittr's functional sequences. I understand if you think assertr is not a right place for these special functions. I can also put them into a separate small package. Maybe as a place to include special functions:

comfirmr::nona(df, var)
comfirmr::isid(df, var)
comfirmr::isdummy(df, var)
comfirmr::positive(df, var)
comfirmr::negative(df, var)
comfirmr::nonzero(df, var)

[tonyfischetti]

(a) thanks for the kind words about the package!
(b) is_id is on the radar! The project is about to be subsumed by ROpenSci so there's a lot of great stuff coming
(c) yeah, assertr might not be the right place for these but they can totally exist in a different package! Before anyone works on that, we should wait to see how the collaboration with ROpenSci will change the way some of those functions are implemented; the API won't change, but there might be changes that will effect the most graceful way to write those convenience functions.
(d) I'm not sure about the namespace thing :( I'll look into it