Changelog

Source: NEWS.md

eesyscreener (development version)

Significant rewrite of the package around a consistent screen_csv() / screen_dfs() orchestration layer, with all checks following the check_*() / precheck_*() naming and return-value conventions.

Performance optimisations including a small file (<5 MB) in-memory screening route.

New checks

• check_col_var_start()
• check_col_var_characteristic()
• check_col_ind_smushed()
• check_filter_ob_total() (migrated from observational_total)
• check_filter_blanks() (migrated from blanks_filters)
• check_geog_ignored_rows() (migrated from ignored_rows)
• check_geog_region_for_la() and check_geog_region_for_lad() (migrated from region_for_la)
• check_geog_na_code() and check_geog_na()
• check_geog_other_dupes() and check_geog_other_code_dupes()
• check_geog_la_col_present(), check_geog_region_col_present(), check_geog_level_completed(), check_geog_overcompleted_cols()
• check_general_null() and check_general_dupes()
• precheck_cross_meta_to_data() and precheck_cross_data_to_meta()
• precheck_filter_not_singular() (migrated from filter_level)
• Data dictionary checks

Infrastructure

• Geography combo checks consolidated into a single check_geog_combos.R with thin per-level wrappers.
• Integration tests added to exercise full CSV screening end-to-end.
• Testthat output no longer dumps unrelated messages.

eesyscreener 0.2.4

• Added check_meta_filter_group_duplicate().
• Added check_meta_geog_catch().
• Added precheck_geog_level() (Hackathon/geographic level).
• Added column snake-case validation (Hackathon/col snake case).
• Implemented indicator exclusion for the filter item limit check.

eesyscreener 0.2.3

• Added dd_check = FALSE option to exclude data dictionary checks from a screening run.
• Cleared out non-accepted test results.

eesyscreener 0.2.2

• Filter default check now handles NA values in the meta column.
• Clean-up around log file handling and associated tests.

eesyscreener 0.2.1

• New filter item limit check.
• Added remaining time checks (Hackathon).
• Added indicator unit validation.
• Added indicator decimal-places validation.
• Added data dictionary column name checks.
• Added stripped filter group check.
• Added whitespace-in-filters check.
• Added check_col_name_spaces() (equivalent to data_variable_spaces).

eesyscreener 0.2.0

• Bugfix: better handling of commas in strings.
• Added indicator grouping checks.
• Added Docker support with precompiled binaries and instructions.
• Added default filter items check.
• Added col_name_spaces and tests (ported from the dashboard).
• Added check_meta_filter_group_is_filter().
• Added precheck_meta_duplicate_label().
• Added result logging.

eesyscreener 0.1.5

• File read-in now fixes all columns to VARCHAR type.

eesyscreener 0.1.4

• Added filter group checks and filter group match function.
• Scan all rows of CSV to guess file types.
• Removed use of ncol() to prevent unnecessary materialisation of data.
• Added check_meta_label().

eesyscreener 0.1.3

• File read-in validation now allows gzipped files as well as CSVs.
• Added filter hint (Hackathon).
• Added parent_filter as a legitimate metadata column.
• Added precheck_col_name_duplicate().
• Added check_meta_indicator_dp().

eesyscreener 0.1.2

• Tidy up DuckDB usage.

eesyscreener 0.1.1

• Increment version number.

eesyscreener 0.1.0

• First tagged release. Adds the API to pkgdown.