Code duplication between test_NA and fix_NA

[davharris]

There's a lot of duplication between the code in test_NA and fix_NA. For example, both functions have lists of NA strings like -999, and it would be easy for these to get out of sync.

What do folks (especially @hilaryparker) think of this division of labor between the functions:

• test_NA finds the missing values
• fix_NA calls test_NA to find where the missing values are, and then replaces them.

If I get the go-ahead, I'm happy to make a pull request for it.

[karthik]

Hi @davharris

Thanks for the note. There is quite likely a lot of code overlap between the test and fix functions. I haven't worried about rebasing code yet, since it feels a bit premature for that. Once all the functions for the first round of tests are complete, it might be worth doing this then.

That said, feel free to go ahead and submit a pull request. Much appreciated. I just can't work on it for another week since there is still hackathon stuff to finish and also a big proposal due next week.

Cheers

[hilaryparker]

Hey @davharris, that sounds like a great idea! Makes it much easier since I was updating both functions at once when adding new NA examples, etc.