Question

我正在将SQL（当前在Python中运行）迁移到Redshift存储过程中，出于性能原因，我希望该proc运行语言列表（例如EN，ES，DE）

看来Redshift存储的过程不支持数组/列表。你们中任何一个可爱的人都知道这样做的最佳方法吗？我已经通过拆分字符串并将其插入到临时表中来工作了，但这有点脏

这不起作用：

CREATE OR REPLACE PROCEDURE sp_del_me(batchid char(36), list_languages char(2)[])
AS $$
BEGIN

  TRUNCATE TABLE control.del_me;

  INSERT INTO control.del_me
  SELECT  sentimentbatchid
  FROM    control.sentiment_batch sb
  WHERE   batchuuid = batchid
  AND     ISOLanguageCode = ANY(list_languages);

END;
$$ LANGUAGE plpgsql;

我收到错误

无效的操作：pg_catalog.bpchar []不是函数或过程支持的参数类型；

这确实有效：

CREATE OR REPLACE PROCEDURE sp_del_me(batchid char(36), list_languages varchar(max))
AS $$
DECLARE
  num_langs INTEGER;
BEGIN

  CREATE TEMP TABLE IF NOT EXISTS tmp_languages(isolanguagecode char(2));
      -- GET THE NUMBER OF LANGUAGES WHICH IS EN OR EN,ES OR EN,ES,DE ETC
      -- SO WE ARE ASSUMING THERE ARE NO GAPS IN THE STRING AND ITS COMMA SEPARATED
      -- I.E. LENGTH OF STRING = 2 + 3 * (1-#LANGUAGES) WHICH, WITH BASIC ALGEBRA EVALUATES TO...
  num_langs := 1 + (len(list_languages) - 2)/3;

  TRUNCATE TABLE tmp_languages;

  FOR counter in 1..num_langs LOOP
    INSERT INTO tmp_languages
    SELECT split_part(list_languages, ',', counter);
  END LOOP;

  TRUNCATE TABLE control.del_me;

  INSERT INTO control.del_me
  SELECT  sentimentbatchid
  FROM    control.sentiment_batch sb
  INNER JOIN tmp_languages tl
    ON sb.isolanguagecode = tl.isolanguagecode
  WHERE   batchuuid = batchid;

END;
$$ LANGUAGE plpgsql;

将数组传递到Redshift存储过程

0 个答案: