What is a schema and how to use it - OpenSearch - Alibaba Cloud Documentation Center

The schema defines the data structure of an OpenSearch application. It has two top-level objects: tables, which describe how your data is organized, and indexes, which control how that data is searched and filtered. Each field in a table must be registered in the schema before it can be searched or used as a filter.

Example

{
    "tables": {
        "main": {
            "primaryTable": true,
            "name": "main",
            "fields": {
                "id": {
                    "name": "id",
                    "type": "LITERAL",
                    "primaryKey": true
                },
                "title": {
                    "name": "title",
                    "type": "TEXT",
                    "primaryKey": false
                },
                "buy": {
                    "name": "buy",
                    "type": "INT",
                    "primaryKey": false
                },
                "cate_id": {
                    "name": "cate_id",
                    "type": "INT",
                    "primaryKey": false
                },
                "cate_name": {
                    "name": "cate_name",
                    "type": "LITERAL",
                    "primaryKey": false
                }
            }
        }
    },
    "indexes": {
        "searchFields": {
            "id": {
                "fields": [
                    "id"
                ]
            },
            "default": {
                "fields": [
                    "title"
                ],
                "analyzer": "chn_standard"
            },
            "cate_name": {
                "fields": [
                    "cate_name"
                ]
            }
        },
        "filterFields": [
            "id",
            "buy",
            "cate_id",
            "cate_name"
        ]
    }
}

This example defines one table (main) with five fields. The indexes object maps three of those fields to search indexes and lists four fields as filter (attribute) fields.

Structure

Schema

Field	Type	Required	Description
`tables`	Object	Yes	Table definitions. Each key is a table name; see Table for the structure of each value.
`indexes`	Object	Yes	Index configuration. See Index.

Table

Example

{
    "primaryTable": true,
    "name": "main",
    "fields": {
        "id": {
            "name": "id",
            "type": "LITERAL",
            "primaryKey": true
        },
        "title": {
            "name": "title",
            "type": "TEXT",
            "primaryKey": false
        }
    }
}

Structure

Field	Type	Required	Description
`name`	String	Yes	The name of the table.
`primaryTable`	Boolean	Yes	Specifies whether this table is the primary table.
`fields`	Object	Yes	Field definitions. Each key is a field name; see Field for the structure of each value.

Field

Example

{
    "name": "json_nested",
    "type": "NESTED",
    "primaryKey": false,
    "innerSchema": {
        "job": {
            "name": "job",
            "type": "TEXT",
            "primaryKey": false
        },
        "ssn": {
            "name": "ssn",
            "type": "LITERAL",
            "primaryKey": false
        }
    }
}

Structure

Field	Type	Required	Description
`name`	String	Yes	The name of the field.
`type`	String	Yes	The field type: `LITERAL`, `TEXT`, `INT`, `NESTED`, or `OBJECT`. For details, see Application schema.
`primaryKey`	Boolean	Yes	Specifies whether this field is the primary key.
`joinWith`	Array	No	Links this field to linked external data tables.
`innerSchema`	Object	No	The nested field structure. Required when `type` is `OBJECT` or `NESTED`. Supports multiple levels of nesting.

Index

Example

{
    "searchFields": {
        "default": {
            "fields": [
                "title"
            ],
            "analyzer": "chn_standard"
        },
        "id": {
            "fields": [
                "id"
            ]
        }
    },
    "filterFields": [
        "id"
    ]
}

Structure

Field	Type	Required	Description
`searchFields`	Object	Yes	Full-text search index definitions. Each key is an index name; see SearchField for the structure of each value.
`filterFields`	Array	Yes	The list of attribute fields.

SearchField

Example

{
    "fields": ["title"],
    "analyzer": "chn_standard"
}

Structure

Field	Type	Required	Description
`fields`	Array	Yes	The table fields included in this search index.
`analyzer`	String	No	The analyzer to apply. For custom analyzers, see Custom analyzers. For built-in analyzers, see Built-in analyzers below and Built-in analyzers.

Built-in analyzers

Chinese

Analyzer	Description
`chn_standard`	General Chinese analysis. Use for most Chinese full-text search fields.
`chn_single`	Single Chinese character analysis. Use when each character is treated as an independent token.
`chn_ecommerce`	Chinese e-commerce analysis. Optimized for product names and shopping queries.
`chn_film`	Chinese video and entertainment analysis. Optimized for movie titles and media content.
`chn_scene_name`	Chinese personal name analysis.
`chn_scene_org`	Chinese organization name analysis.
`chn_it_content`	IT content analysis for Chinese text.

English

Analyzer	Description
`eng_standard`	English analysis with stemming. Matches word variants (e.g., "running" matches "run").
`eng_nostem`	English analysis without stemming. Matches exact word forms only.

Language-agnostic and special-purpose

Analyzer	Description
`simple`	Simple analysis. Splits on whitespace and punctuation, lowercases tokens.
`keyword`	Keyword analysis. Treats the entire field value as a single token. Use for exact-match fields.
`fuzzy`	Fuzzy analysis. Generates token variants to support approximate matching.
`numeric`	Numerical analysis.
`geo`	Geolocation analysis. Use for geographic coordinate fields.
`first_letter`	Simple pinyin spelling analysis. Indexes the first letter of each pinyin syllable.
`full_pinyin`	Full pinyin spelling analysis. Indexes the complete pinyin transcription.