Diagnoses duplicated organization-year filings using the same ranking logic
as deduplicate(). For each duplicated by_id x by_year group, the
function identifies the row that would be retained and classifies each
candidate filing by filing type.
Filing types are:
"normal": not amended, not group, not partial"partial": partial only"amended": amended only"group": group only"amended_partial": amended + partial"amended_group": amended + group"group_partial": group + partial"amended_group_partial": amended + group + partial
The retained filing is indicated separately by the logical column
retained.
Usage
inspect_duplicates(
df,
by_id = "EIN2",
by_year = "TAX_YEAR",
col_group = "RETURN_GROUP_X",
col_partial = "RETURN_PARTIAL_X",
col_amended = "RETURN_AMENDED_X",
col_stamp = "RETURN_TIME_STAMP",
return_candidates = FALSE
)Arguments
- df
A
data.framecontaining IRS 990 e-file data.- by_id
Name of the organization identifier column. Default
"EIN2".- by_year
Name of the tax year column. Default
"TAX_YEAR".- col_group
Name of the group return flag column. Default
"RETURN_GROUP_X".- col_partial
Name of the partial-year return flag column. Default
"RETURN_PARTIAL_X".- col_amended
Name of the amended return flag column. Default
"RETURN_AMENDED_X".- col_stamp
Name of the filing timestamp column. Default
"RETURN_TIME_STAMP".- return_candidates
Logical. If
TRUE, return the full duplicate candidate table with ranking columns. Default isFALSE.
Value
A list with:
- summary_types
Count of duplicated candidate rows by filing type, with total, kept, and dropped counts.
- summary_sizes
Count of duplicated org-years by number of filings.
- summary_size_by_type
Count of duplicated candidate rows by number of filings and filing type, with total, kept, and dropped counts.
- examples
One representative duplicated org-year for each number of filings, including all candidate rows in that example group.
- candidates
Optional full duplicate candidate table, returned only if
return_candidates = TRUE.