Aliyun OSS Vector Store icon

Aliyun OSS Vector Store

Store and retrieve vectors using Alibaba Cloud OSS Vector Bucket

Actions7

List Vectors

AI-generated

Summary

List vectors from a specified Alibaba Cloud OSS vector index with pagination and optional client-side flat metadata filtering.

Inputs

  • Index Name (required) — The name of the OSS vector index to list vectors from; must match an existing index in your bucket.
  • Max Results Per Page — The page size for each ListVectors API call, between 1 and 1000. Lower values reduce the size of each API response.
  • Return Vector Data — Boolean flag indicating whether to include the raw float32 vector data in the response. Default is false to avoid large payloads when only keys or metadata are needed.
  • Return Metadata — Boolean flag to include metadata for each listed vector. This is required for client-side metadata matching and defaults to true.
  • Metadata Match (flat AND) — A JSON object specifying a flat AND filter on metadata fields as string equality. Only vectors whose metadata matches all specified fields are included. Arrays in metadata match if any element equals the value. If empty ({}), lists all vectors with pagination. Use with care on large indexes due to full scan.

Output shape

a list of JSON objects, each containing vector key identifiers and optionally vector data and metadata as specified.

The output includes fields: keys (vector keys listed), deleteKeys (identical to keys, used for deletion), matchedKeyCount (number of keys matching), pagesScanned (number of paginated API calls made), and vectorsListed (total vectors retrieved from OSS). Client-side filtering is applied to metadata per the provided flat AND match. Pagination continues internally until the full listing or filter conditions are met.

Examples

Example 1: List vector keys and metadata for an index with default page size and no vector data

Provide 'Index Name' and leave other options at defaults. The node outputs all vector keys with their metadata.

Example 2: List only vectors whose metadata field 'category' equals 'docs'

Set 'Metadata Match (flat AND)' to {"category":"docs"} to filter client-side for vectors tagged with the 'docs' category.

Links

Discussion