More Like This
Basic Usage
Finds documents similar to a given document or set of field values. This is useful for recommendation engines or finding related content based on textual similarities.
You must pass either:
with_document_id, which takes a “key field” value to match against the corresponding document.with_document_fields, which takes a JSON object string to match against.
All other parameters are compatible with both with_document_id and with_document_fields.
-- with_document_id
SELECT * FROM search_idx.search(
query => paradedb.more_like_this(
with_document_id => 2,
with_min_word_length => 2,
with_max_word_length => 5,
with_boost_factor => 1.0,
with_stop_words => ARRAY['and', 'the', 'for']
)
);
-- with_document_fields
SELECT * FROM search_idx.search(
query => paradedb.more_like_this(
with_document_fields => '{"flavour": "banana"}',
with_min_doc_frequency => 0,
with_max_doc_frequency => 100,
with_min_term_frequency => 1,
)
);
The ID of the document to find similar documents to.
A JSON object representing the field values to use for similarity matching.
Minimum document frequency of terms to be considered.
Maximum document frequency of terms to be considered.
Minimum term frequency of terms to be considered.
Maximum number of query terms to be used.
Minimum word length of terms to be considered.
Maximum word length of terms to be considered.
Boost factor to amplify the impact of matching terms.
A JSON array of stop words to be ignored in the query.
Was this page helpful?