Run the import

The preview button is located next to the Run import button and unlike running the import, the preview does not require an active connection.

When you run a preview, only a sample of the provided data is scanned. This means that the preview can differ from the final import in terms of connectedness, for example. Even so, the preview can be very useful to get an overview of the data, especially on smaller datasets.

Since the preview does not actually import any data, it can be run iteratively until you are satisfied.

You can preview all or selected elements in your data. If your model is particularly complex, it can be beneficial to preview only parts of the data. To use this feature, select the parts of your model that you want to preview and Preview selected from the dropdown.

The preview shows a sample of either all or selected data that is mapped correctly in the model. Unlike the actual import, the preview can be run regardless of the completeness of the mapping. If any element is missing the green checkmark in the model, it will not be included in the preview, but the preview can still be run.

When you are happy with the model and the mapping is complete, the import can begin. Use Run import to start the process.

Verify that the data source is correct, then you need to select which Aura instance to import to. If your instance contains multiple databases, you are prompted to select a target database before running the import.

Depending on your data source, you have two options for import:

Only AWS S3, Azure Blob Storage and Data Lake Storage, and Google Cloud Storage buckets using a non-public authentication method are supported for bulk import.

Once you confirm the operation with Run, you are taken to the Import jobs tab, where you can see the status of the import, and once it’s started, a progress bar. Note that bulk import does not show the progress bar, only the status. The import is done in batches and can be stopped at any time from the […​] menu on the import job.

When the import stops, whether because it is complete or because it was cancelled, you can view a summary from the […​] menu. The summary contains information about the imported nodes and relationships, including time elapsed, the file size, the number of properties etc.

The online import process is idempotent since the LOAD statements use MERGE on ID. This means that if you run the import again, only new data is imported and no existing data is modified nor overwritten. Keep in mind that this is not the case for bulk import, where data is overwritten if you run the import again.

Depending on the size of your data, re-running the import can take some time if you import the whole database again. To optimize this process, you can separate the new/amended data from the original and point Importer to the data source that contains the changes. This way, Importer doesn’t have to parse data that is not going to be imported.

Re-running the online import works the same regardless of data source.