Update index mapping

Update an index mapping

Endpoint - PUT /api/:target/_mapping

Update an index mapping

Request

e.g.

PUT http://localhost:4080/api/myindex/_mapping

Request Body:

{
    "properties": {
        "name": {
            "type": "text",
            "index": true,
            "store": true,
            "sortable": false,
            "aggregatable": false,
            "highlightable": true
        },
        "author": {
            "type": "keyword",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        },
        "price": {
            "type": "numeric",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        },
        "published": {
            "type": "bool",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        },
        "create_at": {
            "type": "date",
            "format": "2006-01-02T15:04:05Z07:00",
            "time_zone": "",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        }
    }
}

Mapping also can defines when create index.

Field types

type	alias	description
text	-	default type, this type will analyze string to some tokens.
keyword	-	use entire string as one token, when query it's just a term.
bool	boolean	like keyword, but the value just true or false
numeric	integer, double, long, short, int, float	numberic type, it supports range query.
date	time, datetime	date type, it supports range query, and date aggregations.

Field type parse rule

text

By default, all string value parse to text type, it can supports match and many queries.

And text type supports two params analyzer and search_analyzer.

analyzer and search_analyzer you should reference to https://zincsearch-docs.zinc.dev/api/index/analyze/

keyword

You should predeined this type.

Usually it use for some special values, it should has limit values, usually it not contians whitespace. like: the days of week, the seasons of year, a tag of something.

it can supports match query, term query, and terms aggregation.

bool

It's very simple, just use for boolean type. it will detected boolean value and set type to bool.

it can supports term query, and terms aggregation.

numeric

All numeric values of integer, double, long, short, int, float will detected and set type to numeric.

it can support term query, terms aggregation, range aggregation and all metrics aggregation, like min, max, avg, sum, count.

date

You should predeined this type. and date type supports param format and time_zone.

format:

date format you can reference to https://pkg.go.dev/time#pkg-constants

Default date format is: 2006-01-02T15:04:05Z07:00, refers to time.RFC3339.

You can set other format what golang support.

Also you can use timestamp for date field, set format to: epoch_millis.

time_zone:

Default value is UTC, you can set it like: +08:00 or -01:00.

Field attributes

All field support 5 attributes:

index

Enable index for the field, default is true, it will can't be query if it disabled.

store

Store value for the field, default is false, it can returns the origin value when query, it used for like highlight.

sortable

Enable sort support for the field, default is false, but it default enabled for numeric and date type.

aggregatable

Enable aggregation for the field, default is false, but it default enabled for numeric and date type.

highlightable

Enable highlight for the field, default is false, if you want use highlight for the field, you should enable it.

Custom @Timestamp field

You can custom `@timestamp`` field in mapping, this allow use your own timestamp.

Default:

{
    "properties": {
        "@timestamp": {
            "type": "date",
            "format": "2006-01-02T15:04:05Z07:00",
            "time_zone": "00:00",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        }
    }
}

Or use custom format and time_zone:

{
    "properties": {
        "@timestamp": {
            "type": "date",
            "format": "2006-01-02 15:04:05",
            "time_zone": "+08:00",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        }
    }
}

Or use epoch_millis:

{
    "properties": {
        "@timestamp": {
            "type": "date",
            "format": "epoch_millis",
            "time_zone": "+08:00",
            "index": true,
            "store": false,
            "sortable": true,
            "aggregatable": true,
            "highlightable": false
        }
    }
}

If use epoch_millis the value of @timestamp should be a numeric of timestamp.

Actually @timestamp is a date field, just named @timestamp.

After defined with mapping, you can use your own time format value for @timestamp field.

Just add the field in you document, like this:

e.g. POST http://localhost:4080/api/myindex/_doc

Request Body:

{
    "name": "Prabhat Sharma",
    "@timestamp": 1656155409850
}

Or use bulk API:

e.g. POST /api/_bulk

{ "index" : { "_index" : "myindex" } } 
{ "name": "Prabhat Sharma", "@timestamp": 1656155409850 }