v0.12.3

• We now support validation checks of Oracle tables via ODBC (#462). (#644, @pachadotdev)

• Redshift table support has now been fixed (#538). (#623, #643, @pachadotdev and @hfrick)

• Added the na_pass argument to col_vals_expr() for finer control of NA values. Previously, NAs were ignored, but now they are caught as failures with the default na_pass = FALSE. As a safeguard, if an expression generates NA values while na_pass is not explicitly supplied, a warning is thrown (#616). (#617)

• Added an infix resolution for log4r_step()'s message argument (#654). (#656, @alexpaynter)

• Fixed the issue where an agent would auto-generating a table label that was too long (truncation occurs now) (#613). (#614)

• Fixed problem where agents would not search the formula environment when materializing ~ tbl (#598). (#599)

• info_columns() now warns more informatively when no columns are selected (#344). (#589).

• write_yaml() errors more informatively now when a tbl value is incompatible for YAML-writing (#596). (#597)

• The yaml_agent_string() function now returns the yaml string (#609). (#610)

• The col_vals_regex() function now supports Perl regexes (#606). (#608)

• We now use a safeguard to fall back to the original string as-is if glue interpolation fails (useful in cases where a user-provided regex like {1,2} is to be preservedin in autobrief but triggers an error by {glue}) (#600). (#601)

• Data extracts for rows_distinct()/rows_complete() now preserve all columns, not just the ones tested (#475). (#588, #591)

• The brief argument of validation functions now also supports {glue} syntax. (#587)

• Validation step briefs correctly recycle to match expanded steps. (#564)

• Performed extensive refactoring of internal dplyr and ggplot2 code. (#579, #582, #583, @olivroy)

• Rebuilt translation table and added support for ordered factors in scan_data(). (#580, @olivroy)

• Modernized test infrastructure and removed deprecated testthat features. (#577, @olivroy)

• Added a new Get Started vignette. (#605, @hfrick)

• The README.md has undergone some refinements. (#618)