What is the safest way to delete all records from DB of a certain project?

[segovend]

As stated in the title, we have a project which had some issues with failing pipelines and it seems it was due to jira JQL that was having 2 shared projects, which one suddenly got archived. Even changing the JQL in jira did not help pipelines still keep failing.
So im wondering how to remove all records even from raw? Is there an existing sql query for this somewhere do to it safely?

Thanks!

[dosubot[bot]]

Hi @segovend! I'm Dosu and I’m helping the incubator-devlake team.

The safest way to delete all records for a specific project—including raw data—in DevLake is to use the built-in API endpoint for project deletion. This endpoint (DELETE /projects/:projectName) calls a service-layer function that handles the removal of the project and all associated data, ensuring dependencies and related records are properly managed. This approach is much safer than running manual SQL queries, as it helps prevent orphaned data and maintains referential integrity throughout the database. You can see the relevant implementation and details here.

There is no official, single SQL script provided for this operation, and direct SQL deletion is discouraged due to the complexity of relationships between tables. If you use the API, make sure to back up your database first and verify that the deletion resolves your pipeline issues.

Note: There is a known bug with Jira data where some records might remain in the domain layer even after deletion, due to how certain raw data tables are referenced. If you suspect lingering Jira data, you may need to manually verify or clean up those tables. More details on this bug and its proposed fix can be found here.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[klesh]

You may delete the scope in the connection page to remove unneeded data.

[segovend]

Hi, thanks. But this means raw still stays, right?
Do I understand correctly that in newer versions you can set the scope meaning what subtasks for collecting data in Jira can be done via GUI?

[klesh]

Raw records should be deleted as well, according to

incubator-devlake/backend/helpers/pluginhelper/api/scope_generic_helper.go

Line 544 in 62b67c1

if strings.HasPrefix(table, "_raw_") {