Vector bucket-level commands provide direct access to the APIs of OSS vector buckets. You can invoke these commands using the vectors-api command. Unlike general-purpose buckets, vector bucket operations do not use object keys. Instead, they manage vector data using indexes and vector keys. All requests must include the --account-id parameter.
A vector bucket API request includes an API name and API parameters, such as the bucket name, index name, header, query, and body.
API name: Each API corresponds to a command. The command name is the hyphen-separated version of the API name.
For example, PutVectorBucket becomes put-vector-bucket. For more information, see API overview.
API parameters
API parameter
Option name
Description
Bucket name
--bucket
string
The name of the bucket.
Index name
--index-name
string
The name of the index. This is required only for vector data operations and index management operations.
Account ID
--account-id
string
The account ID. You must set the account ID in addition to the region. Set the account ID in the configuration file or pass it as a command-line option.
Query parameter
--data-type
string
Each query parameter corresponds to an option. The option name is the parameter name converted from lower camel case to kebab-case, which is hyphen-separated and all lowercase. For example, dataType becomes --data-type.
Body parameter
--vectors
string
These parameters are merged into a single parameter.
The parameter name is the XML root node converted to kebab-case, which is hyphen-separated and all lowercase.
The parameter value is a string. A string prefixed with file:// indicates a file path.
The parameter value also supports the JSON format.