Start Bulk Import — start_import • pineconer

Starts an asynchronous import operation to load vectors from object storage (S3 or GCS) into a Pinecone index.

Usage

start_import(index, uri, integration_id = NULL, error_mode = "continue")

Arguments

index: Name of the index to import into
uri: The URI of the data to import. Supported formats: - S3: "s3://bucket-name/path/to/data/" - GCS: "gs://bucket-name/path/to/data/"
integration_id: Optional ID of the storage integration to use for authentication
error_mode: How to handle errors during import: - "continue": Skip failed records and continue (default) - "abort": Stop the import on first error

Value

List with http response, content (import operation details), and status_code. Content includes: - id: Unique identifier for the import operation - status: Current status (e.g., "Pending", "InProgress")

Details

The data files must be in Parquet format with the following schema: - id (string): Unique vector ID - values (list of floats): Vector values - sparse_values (optional): Sparse vector data - metadata (optional): Key-value metadata

Examples

if (FALSE) { # \dontrun{
# Import from S3
result <- start_import(
  index = "my-index",
  uri = "s3://my-bucket/vectors/"
)

# Check the import ID
import_id <- result$content$id

# Import with error handling
result <- start_import(
  index = "my-index",
  uri = "s3://my-bucket/vectors/",
  error_mode = "abort"
)
} # }