父子模式：如何通过文本/字符串类型的子文档字段搜索文档

Question

Seems that only attribute field can be imported.

Besides, fields of struct can only be attribute indexing mode.

So how can I implement function in vespa like the nested query (text field) in elasticsearch

Answer 1

Parent-child relationships in Vespa are implemented by indexing all parent documents on all nodes in the cluster. These are global documents.

For performance reasons, fields imported from a parent document type must be defined as attribute , which also dictates the type of matching that is supported. Similar restrictions apply for struct fields, regardless of using parent-child relationships. See https://github.com/vespa-engine/vespa/issues/12333 for a feature request on supporting different match modes for imported fields from parent document types.