Check for missing values.

[davharris]

I'm thinking that this would probably be a pretty simple function. By default, it would just return things that were clearly missing (e.g. NA), but it could also (optionally) warn if there were -999s or other common outlier codes.

I'm happy to take this on if nobody else wants it.

Thoughts?

[karthik]

Sure, take it!

[davharris]

Cool. Prototype here. https://github.com/davharris/testdat/blob/master/R/test_NA

Currently outputs a data.frame that identifies the rows and columns with possible NAs, along with the value that triggered inclusion.

> test_NA(dat)
Now checking 2 columns...
  row col   value
1   2   1 missing
2  10   1    <NA>
3   3   2    <NA>

I can make it more robust this afternoon, or change the inputs/outputs as the group suggests.

[karthik]

Mind sending a pull request?