Support skipping conflicting rows during CSV import

[mizukendesu]

When importing CSV data through Supabase Studio, some rows may conflict with existing primary key or unique constraint values.

In cases where a CSV contains both new rows and rows that already exist in the table, it would be helpful if the non-conflicting rows could still be imported while the conflicting rows are skipped.

This would be similar to PostgreSQL's:

INSERT INTO ...
ON CONFLICT DO NOTHING;

I couldn't find an equivalent option in the Studio CSV import flow, so I wanted to suggest it as a possible enhancement.

It would be helpful to have an option in the CSV import UI such as:

• Skip conflicting rows
• Continue importing the remaining valid rows when a primary key or unique constraint conflict occurs
• Report the number of skipped rows after the import completes

Conceptually, this could behave similarly to:

ON CONFLICT DO NOTHING

As a possible future enhancement, users could also be allowed to choose a specific unique constraint as the conflict target. However, simply supporting DO NOTHING on conflicts would already cover the primary use case.

The same behavior can be achieved manually using SQL, scripts, or temporary staging tables.

However, for ad-hoc data migration and operational tasks, having this available directly in Supabase Studio would make CSV imports more convenient.

If there is already a supported way to achieve this in Studio that I have missed, I'd be happy to use that instead.

[paramjeetn]

While Studio's built-in GUI CSV table importer performs atomic batch inserts (failing the transaction on first constraint collision), you can achieve ON CONFLICT DO NOTHING directly using a staging table pattern in the SQL Editor:

Recommended Workflow:

1. Import the CSV into a temporary/staging table (e.g. staging_import with no primary key constraints).
2. Execute the selective merge query:

INSERT INTO target_table (id, name, email, created_at)
SELECT id, name, email, created_at
FROM staging_import
ON CONFLICT (id) DO NOTHING; -- or ON CONFLICT (email) DO NOTHING

3. Verify skipped rows & clean up:

-- See which rows were duplicates/skipped:
SELECT s.* 
FROM staging_import s
LEFT JOIN target_table t ON s.id = t.id
WHERE t.id IS NULL;

-- Drop staging table
DROP TABLE staging_import;

This guarantees that all non-colliding rows are successfully inserted without throwing 23505 duplicate key violations.

[mizukendesu]

Thanks for the detailed workaround.

I’m aware that this can already be handled with a staging table and ON CONFLICT DO NOTHING in SQL, and I mentioned that in the original post as well.

My suggestion is specifically about supporting this behavior directly in the Supabase Studio CSV import UI.

For ad-hoc migrations and operational tasks, having to create a staging table and run a separate merge query adds extra steps for a fairly common case: importing a CSV where most rows are new, but a small number already exist.

Would it make sense to treat this as a Studio feature request?

Even a simple “Skip conflicting rows” option, without initially exposing a configurable conflict target, would cover the main use case.

[mizukendesu]

Supabase team also confirmed to us that this is not currently supported and that our feedback has been added to an existing Dashboard feature request currently in triage.