[Bug]: Failed to get file size for bulk insert API about pymilvus HOT 4 OPEN

This example does a lot of things.
It generates the data files, uploads the files to S3, and then call do_bulk_insert() API to import these files to a milvus server: https://milvus.io/docs/bulk_insert.md#2-Insert-entities

As we mentioned here: https://milvus.io/docs/bulk_insert.md#1-Upload-data-files, to import the files into a milvus server, the data files need to be upload to the bucket defined by minio.bucketName in the milvus.yml configuration file.

From the error message "failed to get file size of 'bulk_data/c08b98aa-cae9-4....." we know the milvus server cannot access this path.
This path is under "a-bucket" bucket. Then I think your milvus server is not using this path as storage path.
Double check the minio.bucketName in the milvus.yml of your milvus server. Change it to be "a-bucket" and try again.

from pymilvus.

/assign @yhmo

from pymilvus.

@yhmo Is it possible to do bulk insert of binary vectors stored as a .npy file?

from pymilvus.

@yhmo Is it possible to do bulk insert of binary vectors stored as a .npy file?

The example script also includes a demo for binary vectors:

from pymilvus.