Skip to content
DuckDB Extensions All extensions Marisa

Query functions in the Marisa DuckDB extension

Function category

Query

3 functions

Membership tests, autocomplete (predictive prefix), and longest-prefix-match against an existing trie BLOB. All three primitives run in O(|key|) — lookup work is independent of dictionary size.

marisa_common_prefix

Scalar function Query
Signature
marisa_common_prefix( col0: BLOB, col1: VARCHAR, col2: INTEGER ) → VARCHAR[]
Arguments (Positional)
Argument col0 Type BLOB Mode Positional Description
Argument col1 Type VARCHAR Mode Positional Description
Argument col2 Type INTEGER Mode Positional Description
Description

Returns VARCHAR[] — every string in the trie that is a prefix of the search string, capped at max_results. The inverse of marisa_predictive: useful for routing tables, IP / CIDR label lookups, and longest-prefix-match style queries ("which rules match this input?").

1 Find every country-code prefix that matches an input 
SELECT marisa_common_prefix(trie, 'USA', 10) AS matches
FROM country_codes_trie;
-- ['U', 'US', 'USA']
Related functions

marisa_lookup

Scalar function Query
Signature
marisa_lookup(col0: BLOB, col1: VARCHAR) → BOOLEAN
Arguments (Positional)
Argument col0 Type BLOB Mode Positional Description
Argument col1 Type VARCHAR Mode Positional Description
Description

Returns BOOLEAN — true iff the search string is in the trie. The fastest membership test in this extension; an exact, allocation-free yes/no against the BLOB. Use it for spell-check, dictionary membership, or anywhere a hash set would otherwise live.

1 Spell-check against an English wordlist trie 
SELECT marisa_lookup(trie, 'duckling') AS in_dictionary
FROM words_trie;
Related functions

marisa_predictive

Scalar function Query
Signature
marisa_predictive( col0: BLOB, col1: VARCHAR, col2: INTEGER ) → VARCHAR[]
Arguments (Positional)
Argument col0 Type BLOB Mode Positional Description
Argument col1 Type VARCHAR Mode Positional Description
Argument col2 Type INTEGER Mode Positional Description
Description

Returns VARCHAR[] — every string in the trie that starts with the given prefix, capped at max_results. The autocomplete primitive: feed it the user's partial input, get back the candidate completions. The cap is a hard cutoff so a short prefix on a million-entry dictionary doesn't return everything.

1 Type-ahead suggestions for a search box 
SELECT marisa_predictive(trie, 'duck', 10) AS suggestions
FROM words_trie;
Related functions