Import Service: Support for List and Vector Datatypes

We are pleased to announce that the Import Service now supports List and Vector datatypes. Whether you are bringing in graph embeddings for GenAI applications or working with list attributes, you can now easily map and import these complex datatypes into Neo4j.

This support is fully integrated across both local file workflows and remote data sources. Here is what you can expect:

For your local CSV workflow you can manage your formatting configurations directly within the files panel. Simply click the menu next to your Data source and select .

This opens a dedicated configuration modal where you can define the exact characters separating your array and vector values.

When importing CSV files hosted in cloud storage buckets, you can configure array and vector delimiters during your data source setup flow. Providing these delimiters triggers the automatic detection of list and vector types, ensuring your cloud-hosted file properties are parsed automatically.

(Parquet, RDBMS, & Cloud Data Warehouses)

If you are connecting to remote data sources that natively support structured or typed data—such as Apache Parquet files, relational databases, or cloud data warehouses—the Data Importer does the heavy lifting for you. Compatible source types are automatically recognized and offered as options mappable directly to their equivalent Neo4j types (such as , , or ).

Once selected, you can easily fine-tune your configuration (such as specifying vector dimensions) right inside the property settings UI.

We are continuously evolving the data loading experience. Please try out the new list and vector support and share your thoughts, feature requests, or bug reports with us at feedback.neo4j.com/data-importer.