<invoke name="query_sql"> <parameter name="sql">WITH r AS (SELECT row_number() OVER () AS id, * FROM raw.iris) SELECT a.species, a.sepal_length AS sl_a, a.sepal_width AS sw_a, a.petal_length AS pl_a, a.petal_width AS pw_a, b.sepal_length AS sl_b, b.sepal_width AS sw_b, b.petal_length AS pl_b, b.petal_width AS pw_b FROM r a JOIN r b ON a.id < b.id AND a.species = b.species AND abs(a.sepal_length-b.sepal_length) <= 0.1 AND abs(a.sepal_width-b.sepal_width) <= 0.1 AND abs(a.petal_length-b.petal_length) <= 0.1 AND abs(a.petal_width-b.petal_width) <= 0.1 ORDER BY a.species, a.sepal_length LIMIT 50</parameter> <parameter name="purpose">Find near-duplicate pairs (same species, all four measurements within 0.1 cm) including the exact duplicate</parameter> </invoke> <invoke_placeholder/>