skipping tests causing cran error, see comments in #205

[geneorama]

I don’t like punting and just skipping the tests, but I don’t know where I’m going to find that special dataset that is not updated, doesn’t risk changing order, and has only NULL values for the first 1,000 rows for at least one field.

And for the second test, I don’t even know what it was testing. It’s also hard to reconstruct without any example. It looks a little familiar and I think we may have developed the example together. I don’t recall though.

The new pull request passes CRAN tests documented in #205

[tomschenkjr]

https://data.cityofchicago.org/Education/Schools-deprecated-2012-/kqmn-byj8

[geneorama]

@tomschenkjr very good example, but I think we need the first 1000 rows to be NA.

Because of #184 the JSON will not return columns that are 100% NA.

I tried ordering the dataset by each column hoping that in one case the first 1000 rows would be 100% NA, but no dice.

Code to reproduce:

names <- c("school_name", "definition", "primary_address", 
           "second_address", "grades", "school2_name", "school2_primary_address", 
           "school2_secondary_address", "school2_grades", "school2_definition", 
           "school3_name", "school3_primary_address", "school3_secondary_address", 
           "school3_grades", "school3_definition", "school4_name", "school4_primary_address", 
           "school4_secondary_address", "school4_grades", "school4_definition")
urls <- sprintf("https://data.cityofchicago.org/resource/kqmn-byj8.json?$order=%s&$limit=1000", names)
dats <- lapply(urls, read.socrata)
lapply(dats, geneorama::NAsummary)
lapply(dats, dim)

[tomschenkjr]

@geneorama Ah! That's right. Thanks for mentioning the underlying issue. I forgot it was completely missing columns on the first page of an API call.

[geneorama]

@nicklucius Thanks for approving (@evejennings FYI)

I'll merge, build, and send to CRAN