find_duplicates: Detect duplicate values

In rmetaverse/synthesisr: Import, Assemble, and Deduplicate Bibliographic Datasets

Detect duplicate values

Description

Identifies duplicate bibliographic entries using different duplicate detection methods.

Usage

find_duplicates(
  data,
  method = "exact",
  group_by,
  threshold,
  to_lower = FALSE,
  rm_punctuation = FALSE
)

Arguments

data	A character vector containing duplicate bibliographic entries.
method	A string indicating how matching should be calculated. Either "exact" for exact matching (the default), or the name of a function for calculating string distance.
group_by	An optional vector, data.frame or list containing data to use as 'grouping' variables; that is, categories within which duplicates should be sought. Defaults to NULL, in which case all entries are compared against all others. Ignored if method = "exact".
threshold	Numeric: the cutoff threshold for deciding if two strings are duplicates. Sensible values depend on the method chosen. Defaults to 5 if method = "string_osa" and must be specified in all other instances except method = "exact" (where no threshold is required).
to_lower	Logical: Should all entries be converted to lower case before calculating string distance? Defaults to FALSE.
rm_punctuation	Logical: Should punctuation should be removed before calculating string distance? Defaults to FALSE.

Value

Returns a vector of duplicate matches, with attributes listing methods used.

See Also

string_ or fuzz_ for suitable functions to pass to methods; extract_unique_references and deduplicate for higher-level functions.

Examples

my_df <-  data.frame(
  title = c(
    "EviAtlas: a tool for visualising evidence synthesis databases",
    "revtools: An R package to support article screening for evidence synthesis",
    "An automated approach to identifying search terms for systematic reviews",
    "Reproducible, flexible and high-throughput data extraction from primary literature",
    "eviatlas:tool for visualizing evidence synthesis databases.",
    "REVTOOLS a package to support article-screening for evidence synthsis"
  ),
  year = c("2019", "2019", "2019", "2019", NA, NA),
  authors = c("Haddaway et al", "Westgate",
              "Grames et al", "Pick et al", NA, NA),
  stringsAsFactors = FALSE
)

# run deduplication
dups <- find_duplicates(
  my_df$title,
  method = "string_osa",
  rm_punctuation = TRUE,
  to_lower = TRUE
)

extract_unique_references(my_df, matches = dups)

# or, in one line:
deduplicate(my_df, "title",
  method = "string_osa",
  rm_punctuation = TRUE,
  to_lower = TRUE)

rmetaverse/synthesisr documentation built on Feb. 23, 2025, 5:29 p.m.