我的团队的观点是建立在观点基础之上的，因此通常DROP TABLE CASCADE是灾难的根源和大量的反复试验。

我想是一个查询，返回了需要在给予一定按照正确的顺序重新创建的所有依赖的对象schema，并table让他们可以自动化和运行了一个脚本。我正在使用RedshiftDROP TABLE文档http://docs.aws.amazon.com/redshift/latest/dg/r_DROP_TABLE.html上的依赖项查询的修改版本。

它似乎是在返回视图及其依赖关系，但不是常规表。我感觉自己接近了，我想念什么？

WITH dependencies AS (
    SELECT DISTINCT
      cls1.oid     AS tbloid,
      nsp1.nspname AS schemaname,
      cls1.relname AS name,
      nsp2.nspname AS refbyschemaname,
      cls2.relname AS refbyname,
      cls2.oid     AS viewoid
    FROM pg_catalog.pg_class cls1
      JOIN pg_catalog.pg_depend dep1
        ON cls1.relfilenode = dep1.refobjid
      JOIN pg_catalog.pg_depend dep2
        ON dep1.objid = dep2.objid
      JOIN pg_catalog.pg_class cls2
        ON dep2.refobjid = cls2.relfilenode
      LEFT OUTER JOIN pg_namespace nsp1
        ON cls1.relnamespace = nsp1.oid
      LEFT OUTER JOIN pg_namespace nsp2
        ON cls2.relnamespace = nsp2.oid
    WHERE dep2.deptype IN ('i' :: "CHAR", 'n' :: "CHAR")
          AND cls2.relkind IN ('v' :: "CHAR", 'r' :: "CHAR")
          AND nsp1.nspname NOT IN ('pg_catalog', 'information_schema')
    ORDER BY 4, 5
),

    joined_to_views AS (
      SELECT
        d.schemaname,
        d.name,
        d.refbyschemaname,
        d.refbyname,
        p.definition
      FROM dependencies d
        LEFT JOIN pg_views p
          ON d.refbyschemaname = p.schemaname AND d.refbyname = p.viewname
  )

SELECT *
FROM joined_to_views