v1.7.1

This release contains several bug fixes, including:

• Users provided an option to handle string-like fields as characters (chr) or factors. Default is to handle string-like fields as character vectors (#27)
• Fixes bug where dates are incorrectly read when first date is a blank (#68)
• Dates are now handled using POSIXct instead of POSIXlt (#8)
• Added additional unit testing (#28)
• Artifacts from Appveyor CI can now be directly submitted to CRAN (#77)
• Fixed issue where JSON may occasionally come back with a final NULL that is not "[]" (in this example it was "[]\n"). This caused getDataFrame to get stuckin an infinite loop while waiting for "[]". Thank you @kevinsmgov for documenting this bug in issue (#96)
• Resolved an error users would see when downloading the JSON file on a SoDA call, Error in rbind(deparse.level, ...) : numbers of columns of arguments do not match. (#19)
• The LIMIT argument was incorrectly interpreted by RSocrata, which sometime causes slower performance (#83, #14)
• The mandatory @context, @id, conformsTo, and describedBy fields were not shown when using ls.socrata(). These elements are now included as attributes in the resulting data frame. (#72)
• Fixed a bug which a user-defined token was ignored when the user passed along a $where argument in a SoDA query. (#105)
• Fixes a bug where the POSIX date conversion would not occur when using the JSON SoDA url (#85)
• Queries now follow SoDA guidelines and are explicitly sorted. This could have caused a silent error where some row may not have downloaded. (#15)

v1.7.0

New features:

• Users can upload data with write.socrata() to upload data to Socrata data portals (using "upsert" and "replace" methods).
• Download private datasets by using Socrata credentials with email and password fields in read.socrata().

Bug fixes:

• Updated unit testing on ls.socrata() to check for @type field is available.
• Converts a Socrata money field into a proper numeric field, instead of a factor.
• Updated build method for Travis to test using the current CRAN packages, not beta packages from GitHub.

v1.6.2

• Resolved potential name collision issue
• Cleaned-up documentation with contributor instructions #23 and #28)
• Moved test coverage in RUnit to testthat and implemented code coverage monitoring (#41)
• Clean-up DESCRIPTION (#40)
• Add continuous integration for Windows (#39)
• Migrate Travis-CI to "proper" R YAML (#46)

v1.6.0

RSocrata v1.6.0 introduces some new functionality and provides some bug fixes. Below is a summary of the changes introduced in this version:

• Added a new function, ls.socrata() to list all datasets on a Socrata portal. Is compatible with Project Open Data Metadata Schema v1.1. For example:

allSitesDataFrame <- ls.socrata("https://soda.demo.socrata.com")
nrow(allSitesDataFrame) # Number of datasets
allSitesDataFrame$title # Names of each dataset

• New optional argument in read.socrata(), app_token, which lets users supply an API token to minimize throttling
• Fixes a bug where read.socrata failed when reading in a date with a column, but there are null values in that column.
• Minor changes to the DESCRIPTION documentation to point users to GitHub for issues and provides new contact information.

v1.5.1

Deprecated httr::guess_media() and implemented httr::guess_type()

v1.4.0

Add json file format for Socrata downloads. Switch to RJSONIO from rjson.

v1.3.0

Added support for human-readable URL. Users can now copy and paste URLs of Socrata-hosted datasets, which will be transformed into a valid SoDA API web query. For example:

earthquakesDataFrame <- read.socrata("https://soda.demo.socrata.com/dataset/USGS-Earthquakes-for-2012-11-01-API-School-Demo/4334-bgaj")

nrow(earthquakesDataFrame) # 1007 (two "pages")

class(earthquakesDataFrame$Datetime[1]) # POSIXlt

Added additional RUnit tests to validate new functionality.

v1.2.0

This release is a minor update that now uses comma-separated file format for Socrata downloads. Available on CRAN.