SQL-MapReduce Call - Aster Analytics

TextChunker requires each sentence to have a unique identifier, and the input to TextChunker must be partitioned by that identifier.

SELECT * FROM TextChunker (
  ON  POSTagger  (
    ON (
      SELECT paraid*1000+sentence_sn
      AS  sentence_id, sentence FROM  Sentenizer  (
        ON paragraphs_input
        TextColumn ('paratext')
        Accumulate ('paraid')
      )
    )
    TextColumn ('sentence')
    Accumulate ('sentence_id')
  ) PARTITION BY  sentence_id  ORDER BY word_sn
  WordColumn ('word')
  POSColumn ('pos_tag')
) ORDER BY 1, 2;