从字符串中每次出现第 n 次后获取文本并对其编号

Posted 2023-03-31

【中文标题】从字符串中每次出现第 n 次后获取文本并对其编号

【英文标题】：Get text after each nth occurrence from string and number them

【发布时间】：2019-08-01 10:22:13

【问题描述】：

通过 Aginity 查询 Redshift

我有一个只有 1 个字段的小表，我想在每次出现 XX 值后获取值，直到下一个空格，还有一列对每次出现进行编号。

MYTABLE：

MYFIELD
The quick XX brown fox XX jumps over the XX lazy dog
Get text XX after each XX nth XX occurrence XX from string

所需的输出：

MYFIELD                                                     OCC   FIELDOUTPUT
The quick XX brown fox XX jumps over the XX lazy dog        1     brown  
The quick XX brown fox XX jumps over the XX lazy dog        2     jumps
The quick XX brown fox XX jumps over the XX lazy dog        3     lazy
Get text XX after each XX nth XX occurrence XX from string  1     after
Get text XX after each XX nth XX occurrence XX from string  2     nth
Get text XX after each XX nth XX occurrence XX from string  3     occurrence 
Get text XX after each XX nth XX occurrence XX from string  4    from

SQL 小提琴：http://sqlfiddle.com/#!15/991c8d

【参考方案1】：

你可以用ORDINALITY分割字符串：

WITH cte AS (
  SELECT *
  FROM MyTABLE, regexp_split_to_table(MYFIELD, E'\\s+') WITH ORDINALITY s(c,rn)
), cte2 AS (
  SELECT myfield, c, LEAD(c) OVER(PARTITION BY MYFIELD ORDER BY rn) AS FieldOutput, rn 
  FROM cte
)
SELECT MYFIELD, Fieldoutput,
       ROW_NUMBER() OVER(PARTITION BY MYFIELD ORDER BY rn) AS occ
FROM cte2
WHERE c = 'XX'
ORDER BY MYFIELD,rn;

db<>fiddle demo

【讨论】：

遇到错误：ERROR: 42601: syntax error at or near "WITH"

@Matt 请尝试使用内联视图的版本：dbfiddle.uk/…

【参考方案2】：

WITH dummy_values AS (
  SELECT 1 UNION ALL 
  SELECT 1 UNION ALL
  SELECT 1
)
, seq AS (
  SELECT (ROW_NUMBER() OVER ())::INT occ 
  FROM dummy_values d1, dummy_values d2, dummy_values d3
)
SELECT 
  "MYFIELD"
, occ
, REGEXP_REPLACE(REGEXP_SUBSTR("MYFIELD", 'XX \\S+', 1, occ), 'XX ', '') fieldoutput
FROM mytable
JOIN seq ON occ <= REGEXP_COUNT("MYFIELD", 'XX ')

【讨论】：

@Matt，此答案对 postgresql 有效，我看到问题已被编辑，现在指的是 redshift，请参阅更新后的答案