Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Duplicated records when row is updated between transactions

Open
  • Is something wrong?
#381 0 comments 1 reaction 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
35/100
Issue type
Bug
Clarity
Mostly clear
Activity status
Stale
Tech stack
cpp, postgresql, python
Domain
databases

Research direction

Start in src/postgres_scanner.cpp around lines 92-93 and inspect why pg_export_snapshot and SET TRANSACTION SNAPSHOT are not appearing in the logged queries. Reproduce with pg_debug_show_queries=true, pg_connection_limit=4, and concurrent PostgreSQL transactions if possible. Done means the snapshot setup occurs as expected and concurrent row updates no longer produce duplicated rows.

Written by the indexing model from the issue text.

Description

What happens?

We started seeing duplicated rows when extracting data from PostgresDB and noticed that it happens when the row is updated/created when different transaction starts, which seems to be similar to what was reported back in https://github.com/duckdb/duckdb-postgres/issues/163.

I checked our logs with the flag pg_debug_show_queries=true after I saw this change and validated all connections are using BEGIN TRANSACTION ISOLATION LEVEL REPEATABLE READ READ ONLY as expected.

However, I noticed SET TRANSACTION SNAPSHOT 'some-snapshot-id' is never called and neither the expected query before that: SELECT pg_is_in_recovery(), pg_export_snapshot(), (select count(*) from pg_stat_wal_receiver).

This query works when executed directly in our database and provides the following result:

SELECT pg_is_in_recovery(), pg_export_snapshot(), (select count(*) from pg_stat_wal_receiver);
 pg_is_in_recovery | pg_export_snapshot  | count 
-------------------+---------------------+-------
 f                 | 0000102C-00400680-1 |     0

I suspect the code is not reaching the part where it gets pg_export_snapshot for some reason and therefore affecting the results of the reading operation even when the isolation level is correct. Our workflow has pg_connection_limit=4.

To Reproduce

A bit hard to reproduce since the update operation of a row needs to happen between the start of concurrent transactions (we see this issue happening a couple of times per week), but we're happy to test some code in our workflows.

OS:

Linux

PostgreSQL Version:

17.5

DuckDB Version:

1.4.1

DuckDB Client:

Python

Full Name:

Rina Sakurai

Affiliation:

Mentimeter

Have you tried this on the latest main branch?
  • I agree
Have you tried the steps to reproduce? Do they include all relevant data and configuration? Does the issue you report still appear there?
  • I agree
Dominant language
C++
Stars
375
Forks
113
Avg merge
6h 42m
Merged PRs (30d)
30

Getting set up

This project ships no dev container, Dockerfile or contributing guide, so setting up is up to you: start from its README, and see our first-contribution guide for the general steps.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from duckdb/duckdb-postgres

All issues in duckdb/duckdb-postgres

Similar issues

More C++ issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.