Question

我有以下查询，我试图删除(语句中第一个CREATE TABLE之后的所有字符以及(以仅显示表名。我有一个查询删除(并清理表名，但我仍然无法在(

之后删除其余字符

redshift_dev=# select stl.userid, trim(pg.usename) AS user,
regexp_replace(
(CASE UPPER(SPLIT_PART(text, ' ', 3))
WHEN 'IF'
THEN SPLIT_PART(text, ' ', 6)
ELSE SPLIT_PART(text, ' ', 3)
END), '(\\()','' ) AS table_name,
starttime as creation_date from stl_ddltext as stl join pg_user as pg on stl.userid = pg.usesysid where text ilike '%create table%' and pg.usename not like '%mcp_etl%' and pg.usename not like '%apiadmin%' and pg.usename not like '%afscloud%' order by creation_date desc limit 1;
 userid | user  |           table_name           |       creation_date
--------+-------+--------------------------------+----------------------------
    100 | admin | analytics.cluster_1000field1 | 2017-10-13 16:17:26.387716
(1 row)

期望的结果：

 userid | user  |           table_name           |       creation_date
--------+-------+--------------------------------+----------------------------
    100 | admin | analytics.cluster_1000 | 2017-10-13 16:17:26.387716
(1 row)

Answer 1

我不确定您的create table stmts通常是什么样的，但您可以尝试使用regexp_substr。

SELECT
  stl.userid
  , trim(pg.usename)        AS user
  , REPLACE(regexp_substr(text, '([\.a-z0-9A-Z]*)\\('), '(', '') as tablename
  , starttime               AS creation_date
FROM stl_ddltext AS stl
  JOIN pg_user AS pg
    ON stl.userid = pg.usesysid
WHERE text ILIKE '%create table%' AND pg.usename NOT LIKE '%mcp_etl%' AND pg.usename NOT LIKE '%apiadmin%' AND pg.usename NOT LIKE '%afscloud%'
ORDER BY creation_date DESC
LIMIT 1;

它基本上尝试在第一个(之前提取所有内容，然后替换(。

从字段中删除剩余字符

1 个答案: