FT.CREATE

Available in:
Redis Stack / Search 1.0.0
Time complexity:
O(K) at creation where K is the number of fields, O(N) if scanning the keyspace is triggered, where N is the number of keys in the keyspace

Create an index with the given specification

Syntax

  FT.CREATE index 
    [ON HASH | JSON] 
    [PREFIX count prefix [prefix ...]] 
    [FILTER {filter}]
    [LANGUAGE default_lang] 
    [LANGUAGE_FIELD lang_attribute] 
    [SCORE default_score] 
    [SCORE_FIELD score_attribute] 
    [PAYLOAD_FIELD payload_attribute] 
    [MAXTEXTFIELDS] 
    [TEMPORARY seconds] 
    [NOOFFSETS] 
    [NOHL] 
    [NOFIELDS] 
    [NOFREQS] 
    [STOPWORDS count [stopword ...]] 
    [SKIPINITIALSCAN]
    SCHEMA field_name [AS alias] TEXT | TAG | NUMERIC | GEO | VECTOR [ SORTABLE [UNF]] 
    [NOINDEX] [ field_name [AS alias] TEXT | TAG | NUMERIC | GEO | VECTOR [ SORTABLE [UNF]] [NOINDEX] ...]

Examples

Required parameters

index

is index name to create. If it exists, the old specification is overwritten.

Optional parameters

ON {data_type}

currently supports HASH (default) and JSON. To index JSON, you must have the RedisJSON module installed.

PREFIX {count} {prefix}

tells the index which keys it should index. You can add several prefixes to index. Because the argument is optional, the default is * (all keys).

FILTER {filter}

is a filter expression with the full RediSearch aggregation expression language. It is possible to use @__key to access the key that was just added/changed. A field can be used to set field name by passing 'FILTER @indexName=="myindexname"'.

LANGUAGE {default_lang}

if set, indicates the default language for documents in the index. Default to English.

LANGUAGE_FIELD {lang_attribute}

is document attribute set as the document language.

A stemmer is used for the supplied language during indexing. If an unsupported language is sent, the command returns an error. The supported languages are Arabic, Basque, Catalan, Danish, Dutch, English, Finnish, French, German, Greek, Hungarian, Indonesian, Irish, Italian, Lithuanian, Nepali, Norwegian, Portuguese, Romanian, Russian, Spanish, Swedish, Tamil, Turkish, and Chinese.

When adding Chinese language documents, set LANGUAGE chinese for the indexer to properly tokenize the terms. If you use the default language, then search terms are extracted based on punctuation characters and whitespace. The Chinese language tokenizer makes use of a segmentation algorithm (via Friso), which segments text and checks it against a predefined dictionary. See Stemming for more information.

SCORE {default_score}

is default score for documents in the index. Default score is 1.0.

SCORE_FIELD {score_attribute}

is document attribute that you use as the document rank based on the user ranking. Ranking must be between 0.0 and 1.0. If not set, the default score is 1.

PAYLOAD_FIELD {payload_attribute}

is document attribute that you use as a binary safe payload string to the document that can be evaluated at query time by a custom scoring function or retrieved to the client.

MAXTEXTFIELDS

forces RediSearch to encode indexes as if there were more than 32 text attributes, which allows you to add additional attributes (beyond 32) using FT.ALTER. For efficiency, RediSearch encodes indexes differently if they are created with less than 32 text attributes.

NOOFFSETS

does not store term offsets for documents. It saves memory, but does not allow exact searches or highlighting. It implies NOHL.

TEMPORARY

creates a lightweight temporary index that expires after a specified period of inactivity. The internal idle timer is reset whenever the index is searched or added to. Because such indexes are lightweight, you can create thousands of such indexes without negative performance implications and, therefore, you should consider using SKIPINITIALSCAN to avoid costly scanning.

When dropped, a temporary index does not delete the hashes as they may have been indexed in several indexes. Adding the DD flag deletes the hashes as well.

NOHL

conserves storage space and memory by disabling highlighting support. If set, the corresponding byte offsets for term positions are not stored. NOHL is also implied by NOOFFSETS.

NOFIELD

does not store attribute bits for each term. It saves memory, but it does not allow filtering by specific attributes.

NOFREQS

avoids saving the term frequencies in the index. It saves memory, but does not allow sorting based on the frequencies of a given term within the document.

STOPWORDS {count}

sets the index with a custom stopword list, to be ignored during indexing and search time. {count} is the number of stopwords, followed by a list of stopword arguments exactly the length of {count}.

If not set, FT.CREATE takes the default list of stopwords. If {count} is set to 0, the index does not have stopwords.

SKIPINITIALSCAN

if set, does not scan and index.

SCHEMA {identifier} AS {attribute} {attribute type} {options...

after the SCHEMA keyword, declares which fields to index:

  • {identifier} for hashes, is a field name within the hash. For JSON, the identifier is a JSON Path expression.

  • AS {attribute} defines the attribute associated to the identifier. For example, you can use this feature to alias a complex JSONPath expression with more memorable (and easier to type) name.

    Field types are:

    • TEXT allows full-text search queries against the value in this attribute.

    • TAG allows exact-match queries, such as categories or primary keys, against the value in this attribute. For more information, see Tag Fields.

    • NUMERIC allows numeric range queries against the value in this attribute. See query syntax docs for details on how to use numeric ranges.

    • GEO allows geographic range queries against the value in this attribute. The value of the attribute must be a string containing a longitude (first) and latitude separated by a comma.

    • VECTOR allows vector similarity queries against the value in this attribute. For more information, see Vector Fields.

    Field options are:

    • SORTABLE: Numeric, tag (not supported with JSON) or text attributes can have the optional SORTABLE argument. As the user sorts the results by the value of this attribute, the results will be available with very low latency. (this adds memory overhead so consider not to declare it on large text attributes).

    • UNF: By default, SORTABLE applies a normalization to the indexed value (characters set to lowercase, removal of diacritics). When using un-normalized form (UNF), you can disable the normalization and keep the original form of the value.

    • NOSTEM: Text attributes can have the NOSTEM argument which will disable stemming when indexing its values. This may be ideal for things like proper names.

    • NOINDEX: Attributes can have the NOINDEX option, which means they will not be indexed. This is useful in conjunction with SORTABLE, to create attributes whose update using PARTIAL will not cause full reindexing of the document. If an attribute has NOINDEX and doesn't have SORTABLE, it will just be ignored by the index.

    • PHONETIC {matcher}: Declaring a text attribute as PHONETIC will perform phonetic matching on it in searches by default. The obligatory {matcher} argument specifies the phonetic algorithm and language used. The following matchers are supported:

      • dm:en - Double metaphone for English
      • dm:fr - Double metaphone for French
      • dm:pt - Double metaphone for Portuguese
      • dm:es - Double metaphone for Spanish

      For more information, see Phonetic Matching.

    • WEIGHT {weight} for TEXT attributes, declares the importance of this attribute when calculating result accuracy. This is a multiplication factor, and defaults to 1 if not specified.

    • SEPARATOR {sep} for TAG attributes, indicates how the text contained in the attribute is to be split into individual tags. The default is ,. The value must be a single character.

    • CASESENSITIVE for TAG attributes, keeps the original letter cases of the tags. If not specified, the characters are converted to lowercase.

    • WITHSUFFIXTRIE for TEXT and TAG attributes, keeps a suffix trie with all terms which match the suffix. It is used to optimize contains (foo) and suffix (*foo) queries. Otherwise, a brute-force search on the trie is performed. If suffix trie exists for some fields, these queries will be disabled for other fields.

Notes:

  • Attribute number limits: RediSearch supports up to 1024 attributes per schema, out of which at most 128 can be TEXT attributes. On 32 bit builds, at most 64 attributes can be TEXT attributes. The more attributes you have, the larger your index, as each additional 8 attributes require one extra byte per index record to encode. You can always use the NOFIELDS option and not encode attribute information into the index, for saving space, if you do not need filtering by text attributes. This will still allow filtering by numeric and geo attributes.

  • Running in clustered databases: When having several indices in a clustered database, you need to make sure the documents you want to index reside on the same shard as the index. You can achieve this by having your documents tagged by the index name.

       127.0.0.1:6379> HSET doc:1{idx} ...
       127.0.0.1:6379> FT.CREATE idx ... PREFIX 1 doc: ...
       

    When Running RediSearch in a clustered database, you can span the index across shards using RSCoordinator. In this case the above does not apply.

Return

FT.CREATE returns a simple string reply OK if executed correctly, or an error reply otherwise.

Examples

Create an index

Create an index that stores the title, publication date, and categories of blog post hashes whose keys start with blog:post: (for example, blog:post:1).

127.0.0.1:6379> FT.CREATE idx ON HASH PREFIX 1 blog:post: SCHEMA title TEXT SORTABLE published_at NUMERIC SORTABLE category TAG SORTABLE
OK

Index the sku attribute from a hash as both a TAG and as TEXT:

127.0.0.1:6379> FT.CREATE idx ON HASH PREFIX 1 blog:post: SCHEMA sku AS sku_text TEXT sku AS sku_tag TAG SORTABLE

Index two different hashes, one containing author data and one containing books, in the same index:

127.0.0.1:6379> FT.CREATE author-books-idx ON HASH PREFIX 2 author:details: book:details: SCHEMA
author_id TAG SORTABLE author_ids TAG title TEXT name TEXT

In this example, keys for author data use the key pattern author:details:<id> while keys for book data use the pattern book:details:<id>.

Index a JSON document using a JSON Path expression

Index authors whose names start with G.

127.0.0.1:6379> FT.CREATE g-authors-idx ON HASH PREFIX 1 author:details FILTER 'startswith(@name, "G")' SCHEMA name TEXT

Index only books that have a subtitle.

127.0.0.1:6379> FT.CREATE subtitled-books-idx ON HASH PREFIX 1 book:details FILTER '@subtitle != ""' SCHEMA title TEXT

Index books that have a "categories" attribute where each category is separated by a ; character.

127.0.0.1:6379> FT.CREATE books-idx ON HASH PREFIX 1 book:details FILTER SCHEMA title TEXT categories TAG SEPARATOR ";"

Index a JSON document using a JSON Path expression.

127.0.0.1:6379> FT.CREATE idx ON JSON SCHEMA $.title AS title TEXT $.categories AS categories TAG

See also

FT.ALTER

History

  • Starting with Redis version 2.0.0: Added PAYLOAD_FIELD argument for backward support of FT.SEARCH deprecated WITHPAYLOADS argument
  • Starting with Redis version 2.0.0: Deprecated PAYLOAD_FIELD argument

Feedback

If you've found issues on this page, or have suggestions for improvement, please submit a request to merge or open an issue in the repository.