🧚 Écoutez ! Beekeeper Studio est une interface de base de données rapide, moderne et open source Télécharger
January 7, 2026 Par Matthew Rathbone

Let me share my experience with handling duplicate data in SQL databases. I’ve dealt with this issue many times, and I’ll show you some practical ways to clean up those pesky duplicates!

Throughout this guide, you’ll find interactive SQL exercises where you can practice identifying and handling duplicate records directly in your browser. These hands-on examples will help you master using DISTINCT, GROUP BY with HAVING, and ROW_NUMBER() to clean up your data effectively.

🔖 Quick Reference: Check out our SQL Language Cheat Sheet for quick lookups on SELECT DISTINCT, GROUP BY, HAVING, and other essential SQL commands for data cleaning.

If you’re looking for a user-friendly SQL editor to practice these techniques, Beekeeper Studio’s SQL editor provides helpful features like syntax highlighting and query execution.

Why Duplicates Are a Headache

Let me tell you why I take duplicate data seriously. Imagine having multiple copies of the same customer record in your database – it’s like having multiple copies of the same contact in your phone! Here’s why this drives me crazy:

  • It messes up your reports: When I’m trying to count unique customers, duplicates make my numbers incorrect
  • It slows everything down: My queries take longer to run because they’re processing the same data multiple times
  • It wastes storage space: Why pay for extra storage when you’re just storing the same information repeatedly?

How I Spot Duplicates

Before I clean up duplicates, I need to find them first. Here are my go-to methods:

The Simple Way: Using DISTINCT

This is the easiest method I use when I just want to see unique values:

SELECT DISTINCT name, email, phone
FROM customers;

Try it yourself:

[[ expanded ? '▼' : '▶' ]]

[[ testData.title ]]

Query the customers table to return only unique combinations of name and email. The table has duplicate entries that need to be identified.

Exemple Interactif ✓ Terminé
Colonnes attendues : [[ col ]]
ℹ️ Cet exercice réinitialise la base de données à chaque exécution. Écrivez votre solution complète en une seule soumission.
[[ i18n.correct ]] [[ validationResult ? i18n.all_checks_passed : i18n.query_success ]]
[[ detail.message ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.not_quite_right ]] [[ patternError ]] [[ validationResult.error ]] [[ i18n.results_dont_match ]]
[[ detail.passed ? '✓' : '✗' ]] [[ detail.message ]]
[[ i18n.your_results ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.expected_results ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.sql_error ]]
[[ error ]]
💡 [[ i18n.hint_label ]] [[ testData.hint ]]
📊 [[ i18n.expected_result_label ]]
[[ col ]]
[[ formatCell(cell) ]]
✨ [[ i18n.solution_label ]]
[[ testData.solution ]]
Tables Disponibles
[[ table.name ]]
[[ col ]]
[[ formatCell(cell) ]]

The Detective Way: Finding Exact Duplicates

When I want to actually see which records are duplicated:

SELECT name, email, phone, COUNT(*) as duplicate_count
FROM customers
GROUP BY name, email, phone
HAVING COUNT(*) > 1;

Try it yourself:

[[ expanded ? '▼' : '▶' ]]

[[ testData.title ]]

Query the products table to find which product_name/category combinations appear more than once. Show the duplicated values and how many times they appear.

Exemple Interactif ✓ Terminé
Colonnes attendues : [[ col ]]
ℹ️ Cet exercice réinitialise la base de données à chaque exécution. Écrivez votre solution complète en une seule soumission.
[[ i18n.correct ]] [[ validationResult ? i18n.all_checks_passed : i18n.query_success ]]
[[ detail.message ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.not_quite_right ]] [[ patternError ]] [[ validationResult.error ]] [[ i18n.results_dont_match ]]
[[ detail.passed ? '✓' : '✗' ]] [[ detail.message ]]
[[ i18n.your_results ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.expected_results ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.sql_error ]]
[[ error ]]
💡 [[ i18n.hint_label ]] [[ testData.hint ]]
📊 [[ i18n.expected_result_label ]]
[[ col ]]
[[ formatCell(cell) ]]
✨ [[ i18n.solution_label ]]
[[ testData.solution ]]
Tables Disponibles
[[ table.name ]]
[[ col ]]
[[ formatCell(cell) ]]

My Favorite Ways to Remove Duplicates

Let me share the methods I use most often to clean up duplicate data:

1. The Quick Way: Using ROW_NUMBER()

This is my favorite method because it’s clean and efficient:

WITH DuplicateCTE AS (
    SELECT *,
        ROW_NUMBER() OVER (
            PARTITION BY name, email, phone
            ORDER BY id
        ) as row_num
    FROM customers
)
DELETE FROM DuplicateCTE
WHERE row_num > 1;

Try it yourself:

[[ expanded ? '▼' : '▶' ]]

[[ testData.title ]]

Query the transactions table using ROW_NUMBER() to identify which duplicate transactions should be kept (row_num = 1) and which should be removed (row_num > 1). Show all records with their row numbers.

Exemple Interactif ✓ Terminé
Colonnes attendues : [[ col ]]
ℹ️ Cet exercice réinitialise la base de données à chaque exécution. Écrivez votre solution complète en une seule soumission.
[[ i18n.correct ]] [[ validationResult ? i18n.all_checks_passed : i18n.query_success ]]
[[ detail.message ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.not_quite_right ]] [[ patternError ]] [[ validationResult.error ]] [[ i18n.results_dont_match ]]
[[ detail.passed ? '✓' : '✗' ]] [[ detail.message ]]
[[ i18n.your_results ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.expected_results ]]
[[ col ]]
[[ formatCell(cell) ]]
[[ i18n.sql_error ]]
[[ error ]]
💡 [[ i18n.hint_label ]] [[ testData.hint ]]
📊 [[ i18n.expected_result_label ]]
[[ col ]]
[[ formatCell(cell) ]]
✨ [[ i18n.solution_label ]]
[[ testData.solution ]]
Tables Disponibles
[[ table.name ]]
[[ col ]]
[[ formatCell(cell) ]]

This interactive example shows how ROW_NUMBER() identifies which records to keep (row_num = 1) versus which would be removed (row_num > 1). In production, you’d then delete where row_num > 1.

2. The Safe Way: Creating a Clean Table

When I want to be extra careful with important data:

-- First, I create a new table with only unique records
SELECT DISTINCT name, email, phone
INTO customers_clean
FROM customers;

-- Then, if everything looks good, I can replace the old table
DROP TABLE customers;
RENAME TABLE customers_clean TO customers;

Tips From My Experience

Here’s what I’ve learned from cleaning up lots of duplicate data:

  • Always backup first: I can’t stress this enough – make a backup before deleting anything!
  • Start small: I always test my duplicate removal queries on a small subset of data first
  • Check your work: After removing duplicates, I double-check that I kept the right records
  • Prevent future duplicates: I add unique constraints to prevent duplicates from coming back

Keeping Your Data Clean

Here’s how I prevent duplicates from happening in the first place:

  1. Use primary keys: I always add a unique identifier for each record
  2. Add unique constraints: I put these on columns that should never have duplicates
  3. Regular checks: I run duplicate checks regularly to catch issues early

Remember, cleaning up duplicates might seem scary at first, but with these methods, it becomes much more manageable. The key is to be careful, test thoroughly, and always have a backup plan!

Beekeeper Studio Est Une Interface de Base de Données Gratuite et Open Source

Le meilleur outil de requêtes SQL et éditeur que j'ai jamais utilisé. Il fournit tout ce dont j'ai besoin pour gérer ma base de données. - ⭐⭐⭐⭐⭐ Mit

Beekeeper Studio est rapide, intuitif et facile à utiliser. Beekeeper prend en charge de nombreuses bases de données et fonctionne très bien sur Windows, Mac et Linux.

La version Linux de Beekeeper est 100% complète, sans compromis sur les fonctionnalités.

Ce Que Les Utilisateurs Disent De Beekeeper Studio

★★★★★
"Beekeeper Studio a complètement remplacé mon ancien workflow SQL. C'est rapide, intuitif et rend le travail avec les bases de données agréable à nouveau."
— Alex K., Développeur de Bases de Données
★★★★★
"J'ai essayé de nombreuses interfaces de bases de données, mais Beekeeper trouve l'équilibre parfait entre fonctionnalités et simplicité. Ça marche tout simplement."
— Sarah M., Ingénieure Full Stack