Meet "Connection xxx not found" when exporting a project

Haoran

Thanks for your time at the beginning.

I am currently exporting a project, only check the 4 default options.

However, it failed with a warning "An invalid argument has been encountered : Connection 'SF_VAW_PROD_ATP_MED' does not exist"Then I tried to use Python API to find this connection, but failed again:

import dataiku




def get_project(client, project_key):

    return client.get_project(project_key)


def find_connection(project, target_connection) -> str:

    for dataset in project.list_datasets(as_type='objects'):

        ds_settings = dataset.get_settings()

        current_connection = ds_settings.get_raw().get('params').get('connection')

        if current_connection == target_connection:

            return dataset.name

    return "not found"




if __name__ == "__main__":


    # Dataiku

    url = "https://xxxxxxxxxxxxxxxxxxxxx"

    api_key = "dkuaps-xxxxxxxxxxxxxxxxxxx"

    target_connection = "MY_CONNECTION"


    # set up client

    dataiku.set_remote_dss(url, api_key, no_check_certificate=True)

    client = dataiku.api_client()


    # project key

    project_key = "MY_PROJECT"

    print("project_key: " + project_key)


    # project

    project = client.get_project(project_key)


    # function call

    dataset_name = find_connection(project, target_connection)


    print("Dataset: " + dataset_name)

===========================================================================================


Result:

project_key: MY_PROJECT
Dataset: not found

Do you have any ideas? Appreciate your help a lot.

Operating system used: Windows 11 Enterprise

Alexandru

Hi,
It appears that the connection 'SF_VAW_PROD_ATP_MED' has been deleted, and some datasets are still referencing it, so the export fails.

To overcome this, you temporarily re-create the connection ( it doesn't have to be working, don't add any credentials, just have the same name)
Next, determine which datasets are still referencing this connection. Per-connection data in (admin/monitoring/connection-data/) changes the connection for the datasets from the flow; re-export the project.
You can finally delete the connection after all objects have been moved to a new connection.

Thanks

Alexandru

Additionally there may be other object beyond datasets that are using the connection like SQL Noteboooks which you may need to check