16.20 - nPath Resultの例2: FIRSTおよびACCUMULATEの3つの形式 - Teradata Vantage NewSQL Engine

Teradata Vantage™ - NewSQL Engineの分析関数

prodname
Teradata Database
Teradata Vantage NewSQL Engine
vrm_release
16.20
category
プログラミング リファレンス
featnum
B035-1206-162K-JPN

入力

clicks
userid sessionid productname pagetype clicktime referrer productprice
1039 1 NULL home 06:59:13 Company1 100
1039 1 NULL home 07:00:10 Company2 300
1039 1 television checkout 07:00:12 Company2 500
1039 1 television checkout 07:00:18 Company2 10
1039 1 envelopes checkout 07:01:00 Company3 10
1039 1 NULL checkout 07:01:10 Company3 10

SQL呼び出し

SELECT * FROM nPath (
  ON clicks PARTITION BY sessionid ORDER BY clicktime
  USING
  Mode (NONOVERLAPPING)
  Symbols (
    pagetype='home' AS H,
    pagetype='checkout' AS C,
    pagetype <> 'home' AND pagetype <>'checkout' AS A
  )
  Pattern ('^H+.A*.C+$')
  Result (
    FIRST (sessionid OF ANY (H, A, C)) AS sessionid,
    FIRST (clicktime OF H) AS firsthome,
    FIRST (clicktime OF C) AS firstcheckout,
    ACCUMULATE (productname OF ANY (H,A,C) DELIMITER '*') AS products_accumulate,
    ACCUMULATE (CDISTINCT productname OF ANY (H,A,C) DELIMITER '$$') AS cde_dup_products,
    ACCUMULATE (DISTINCT productname OF ANY (H,A,C)) AS de_dup_products
  )
) ORDER BY sessionid;

出力

sessionid firsthome firstcheckout products_accumulate cde_dup_products de_dup_products
1 06:59:13 07:00:12 [null*null*television*television*envelopes*null] [null$$television$$envelopes$$null] [null, television, envelopes]