sc_select() function|>update dictionary for 12 December 2019 release of data
users can now request field of study-level data elements
updates to sc_dict() command:
return_df argumentprint_off argumentprint_dev == TRUE argumentprint_notes == TRUEusers can now request JSON version of returned data by using
return_json == TRUE in sc_get()
returned data frame now puts variables in order requested in
sc_select()
some new tests
sc_get (h/t @nguyentr17)utility_functions::dev_to_varsc_year() is now 'latest' rather
than 2013. With continued data updates, this makes more sense than
keeping an old year. Existing scripts that relied on the default for
data from 2013 will need to be updated.year column will be a character column with
latest as the value when the most recent data are choosen. The
College Scorecard doesn't clearly note which data are the latest, so
I have left the string. When building a panel dataset across
multiple years, it will be best to use numeric year values for all
years so that the resulting tibbles can be bound together cleanly.sc_select(): starts_with(),
ends_with(), contains(), and matches() should now be
available.sc_select_() and sc_filter_(), which allow users to select
and filter variables using strings stored in environment variablesc_zip() to take zip codes that start with zero (h/t
@nateaff), either with string value or by returning leading zeros to
numeric values that R dropshttr to make call rather than jsonlite directly) in order to improve parsing on bad linesdebug option to sc_get() so that the API URL string could be returned when debugging callsysdata.rda in ./data-raw/make_dict_hash.Rsc_get() to use floor() instead of ceiling() so that
it doesn't make unnecessary API request/pull (h/t @jjchern)sc_filter() to use subset object vectorssc_filter() to use vectors stored in objectssc_filter() to use %in% operatorsc_dict() to search all columns by defaultsc_dict() bug that wouldn't allow for search by developer friendly names