Why can't I select Id in a GROUP BY query?

Every selected field in a grouped query must be either a grouping column or an aggregate. `Id` is unique per row, so grouping on it would produce one group per record.

Does this replace Salesforce duplicate rules for accounts?

No. Duplicate rules and matching rules prevent new duplicates at save time with fuzzy matching. This query is exact-match, and it is for auditing the records already in the org.

How do I catch duplicates that differ only in case?

SOQL grouping is case-insensitive for text fields in most orgs, but leading/trailing whitespace still splits groups. Export and normalise in a spreadsheet when the match must be fuzzy.

SOQL Query to Find Duplicate Accounts by Name

Find duplicate accounts in Salesforce with a GROUP BY / HAVING SOQL query on Name, then pull the underlying record IDs.

Updated 2026-09-09

Duplicate detection in SOQL is a two-step job: GROUP BY ... HAVING COUNT(Id) > 1 to find the offending values, then a second query to pull the records behind them.

Accounts sit at the top of most data models, so filters here usually cascade into contact, opportunity, and case reporting.

Step 1 - find the duplicated values

HAVING filters the grouped result; WHERE filters rows before grouping. You will usually want both - WHERE to scope the time window, HAVING to keep only the collisions.

Duplicated Name values

SELECT Name, COUNT(Id) total
FROM Account
GROUP BY Name
HAVING COUNT(Id) > 1
ORDER BY COUNT(Id) DESC

Step 2 - pull the records behind a duplicate

Aggregate queries do not return record IDs. Take a value from step 1 and query the records so you can decide which one to keep.

The colliding records

SELECT Id, Name, Type, Industry, OwnerId
FROM Account
WHERE Name = 'paste-value-here'
ORDER BY CreatedDate

Scope it to recent data

On a large object, group across the whole table only once. After the initial cleanup, scope to a rolling window so the query stays fast and the list stays actionable.

Last 90 days only

SELECT Name, COUNT(Id) total
FROM Account
WHERE CreatedDate = LAST_N_DAYS:90
GROUP BY Name
HAVING COUNT(Id) > 1

Common mistakes to avoid

Use the API name Account, not the UI label - the two differ on several standard objects.
Wrap literal text values in single quotes; date literals such as LAST_N_DAYS:30 must stay unquoted.
Run the query in a sandbox first if it feeds a bulk update - a WHERE clause that is one character off can select the whole table.
GROUP BY cannot be combined with Id in the SELECT list - that is the error most people hit first.
Case and whitespace differences make near-duplicates invisible to GROUP BY; a trailing space is a different value.

Run it faster with TurboKit

Deduplication is an export-edit-reimport loop, and the export half is where most of the time goes.

TurboKit's AI SOQL builder lets you describe the result you want in plain English, returns the query, and exports the rows to CSV or Excel in the same panel - without leaving the Salesforce tab you are already on.

Frequently asked questions

Why can't I select Id in a GROUP BY query?: Every selected field in a grouped query must be either a grouping column or an aggregate. Id is unique per row, so grouping on it would produce one group per record.
Does this replace Salesforce duplicate rules for accounts?: No. Duplicate rules and matching rules prevent new duplicates at save time with fuzzy matching. This query is exact-match, and it is for auditing the records already in the org.
How do I catch duplicates that differ only in case?: SOQL grouping is case-insensitive for text fields in most orgs, but leading/trailing whitespace still splits groups. Export and normalise in a spreadsheet when the match must be fuzzy.