• Add progress bar when downloading extracts
  • Removed validate argument from extract revision functions and improved warning messages for invalid extract field names when used as arguments.
  • Fixed bug preventing users from providing their API key directly to submit_extract and wait_for_extract.
  • Added the “Rmd for Reproducible Research” template, which sets up a workflow that leverages the IPUMS microdata extract API to facilitate sharing your analysis. For more details, see the blog post. Credit to @ehrlichd for the template and blog post!

  • Moved the raster package to Suggests so that it is no longer installed automatically when you install ipumsr. The raster package is only required if you need to read raster extracts from the IPUMS Terra collection, and IPUMS Terra is slated to be decommissioned shortly.

  • Fixed bug causing a read error for some labeled string variables (#61, thanks @chengchou).

  • ipumsr now always uses the haven::labelled() function to create labelled vectors, in order to maintain compatibility with developments in the haven and vctrs packages (thanks @gergness!).

  • ipumsr now requires R 3.5 or greater, in line with the new requirements of package dependency raster.

  • Modify lbl_define() test to reflect changes to haven’s labelled class definition.
  • Add lbl_define() function to enable the use of lbl_relabel() syntax when
    creating a new labelled vector from an unlabelled one (#51, thanks @chengchou).

  • Remove pillar printing from ipumsr, getting rid of pesky warning (#47).

  • Improved documentation for lower_vars argument (#56, thanks @hrecht).

  • Incorporate bug fix in knitr 1.23 that affected encoding in NHGIS vignette.
  • Remove stringr & tidyr dependencies so installation is a little easier (#41).

  • Fix bug in pillar printing of haven’s labelled objects (#43)

  • Lots of improvements for users who wish to use “big data” sized IPUMS extracts. See the vignette using command vignette("ipums-bigdata", package = "ipusmr") for the full details.

  • Added pillar printing for labelled classes in tibbles. This means that the label will print the labels alongside the values when printed in a tibble (in a subtle grey color when the terminal supports it). To turn this feature off, use command `options(“ipumsr.show_pillar_labels” = FALSE).

  • The approach to reading hierarchical data files is much faster.

  • Arguments to read_ipums_sp() are now in the same order as read_ipums_sf()

  • read_ipums_sf() and read_ipums_sp() gain 2 new arguments vars which allows you to pick a subset of variables, and add_layer_var which lets you add a variable indicating which layer it came from.

  • You can now use your inside voice for variable names with the new argument lower_vars for read_ipums_ddi() and read_ipums_micro() family of functions so that the variable names are lower case.

  • ipumsr is compatible with versions of haven newer than 2.0 (while maintaining compatibility with earlier versions). (#31)

  • IPUMS Terra is now officially supported! Read raster, area or microdata extracts using functions read_terra_raster(), read_terra_raster_list(), read_terra_area(), read_terra_area_sf(), and read_terra_micro()

  • Add support for keyvar in DDI, which will (eventually) help link data across record types in hierarchical extracts. To be effective, this requires more support on the ipums.org website, which is hopefully coming soon (#25 - thanks @mpadge!)

  • Improved main vignette instructions for Safari users (#27)

  • Fix for selecting columns from csv extracts (#26 - thanks forum user JCambon_OIS!)

  • Fixes for the ipums_list_*() family of functions.

  • Fixed a bug in ipums_shape_*_join functions when using integer ID columns. (#16)

  • Allow for unzipped folders because Safari on macOS unzips folders by default (#17)

  • lbl_relabel behavior is improved so that labels aren’t assigned sequentially (#21)