Changes in version 0.8.0 (2026-03-31)                  

Multi-Backend Join Support

  - Join wrappers (left_join_spy(), join_strict(), etc.) now auto-detect
    input class and dispatch to the native join engine: tibble inputs
    use dplyr, data.table inputs use data.table, plain data.frames use
    base R merge()
  - New backend parameter for explicit override ("base", "dplyr",
    "data.table")
  - Class preserved through the full diagnose-repair-join cycle
  - Graceful fallback to base R with a warning if the detected backend
    package is not installed

Documentation Overhaul

  - README reframed around string-level key diagnostics (whitespace,
    case, encoding, typos) as the primary value proposition
  - Three new vignettes: "Why Your Keys Don't Match" (real-world
    scenarios), "Joins in Production" (pipelines, logging, guards),
    "Working with Backends" (tibble/data.table/base R)
  - Existing vignettes (quickstart, common-issues) expanded from 850
    to 4,500+ prose words

Testing

  - 46 new backend tests covering auto-detection, explicit override,
    named by vectors, class preservation, and all three backends

                        Changes in version 0.7.2                        

Testing & Quality

  - Improved test coverage from 65% to 94%
  - Added comprehensive tests for all exported and internal functions
  - Added .covrignore to exclude untestable Shiny addin from coverage
    calculation

                        Changes in version 0.7.1                        

API Simplification

  - Removed plot_venn(): Use plot(report, file = "...") instead
  - Removed plot_summary(): Use summary(report, format = "markdown")
    instead
  - Enhanced plot(): Now accepts file, width, height, colors parameters
  - Enhanced summary(): Now accepts format parameter ("data.frame",
    "text", "markdown")

CRAN Preparation

  - Added cph role and ORCID to DESCRIPTION
  - Added Language: en-US field
  - Quoted proper nouns in documentation

                        Changes in version 0.7.0                        

Integration & Workflow

  - RStudio addin: addin_join_inspector() opens an interactive Shiny
    gadget for exploring join diagnostics
  - Logging & audit: log_report() writes reports to file (text/JSON/RDS
    format)
  - Automatic logging: set_log_file() / get_log_file() for session-wide
    logging configuration
  - Deferred access: last_report() retrieves diagnostics after silent
    joins
  - Silent mode: .quiet = TRUE parameter for all *_join_spy() wrappers

                        Changes in version 0.6.0                        

Advanced Join Patterns

  - check_cartesian(): Detect and warn about cross-join / Cartesian
    product explosions
  - detect_cardinality(): Determine actual key relationship (1:1, 1:n,
    n:1, n:m)
  - analyze_join_chain(): Analyze cascading joins in multi-table
    workflows

                        Changes in version 0.5.0                        

Visualization

  - plot_venn(): Venn diagram of key overlap with PNG/SVG/PDF export
  - plot_summary(): Summary tables in text, markdown, or data.frame
    format
  - S3 integration: plot() method for JoinReport objects

                        Changes in version 0.4.0                        

Performance & Scale

  - Sampling mode: join_spy(x, y, by, sample = 10000) for large datasets
  - Memory estimation: Predict join result size before execution
  - Reproducible sampling: Optional seed parameter for consistent
    results

                        Changes in version 0.3.0                        

Auto-Repair

  - join_repair(): Automatically fix common key issues
      - Trim leading/trailing whitespace
      - Standardize case (to lower/upper)
      - Remove invisible characters
      - Convert empty strings to NA
  - suggest_repairs(): Generate R code snippets to fix detected issues
  - Dry-run mode: Preview changes without modifying data

                        Changes in version 0.2.0                        

Enhanced Detection

Type & Format Issues

  - Type coercion warnings (character to factor, numeric-as-string to
    numeric)
  - Factor level mismatches between tables
  - Empty string vs NA distinction
  - Numeric precision warnings for floating-point keys

Near-Match Detection

  - Fuzzy match candidates (Levenshtein distance ≤ 2)
  - Suggested matches for unmatched keys
  - Configurable similarity threshold

Multi-Column Join Analysis

  - Per-column breakdown for composite keys
  - Problem column identification (lowest match rate)

                        Changes in version 0.1.0                        

Initial release.

Core Functions

  - join_spy() - Comprehensive pre-join diagnostic report
  - key_check() - Quick key quality assessment
  - key_duplicates() - Find and locate duplicate keys
  - join_explain() - Explain row count changes post-join
  - join_diff() - Before/after state comparison
  - join_strict() - Join with cardinality enforcement
  - left_join_spy(), right_join_spy(), inner_join_spy(), full_join_spy()
    - Join wrappers with diagnostics

Issue Detection

  - Duplicate key detection with row locations
  - Whitespace detection (leading/trailing spaces)
  - Case mismatch detection
  - Encoding issue detection (UTF-8 vs Latin-1, invisible Unicode)
  - NA key detection
  - Cardinality analysis (1:1, 1:n, n:1, n:m)

S3 Methods

  - print(), summary(), plot() methods for JoinReport objects

Documentation

  - Getting started vignette
  - Common issues troubleshooting guide