Question

请告诉我是否有可用于REGEXP_LIKE的替代功能？

我有以下情况：

Names:
ID NAME
1  Alex
2  Jim
3  Tom

Invoices:
ID Amount Name_IDs
1  100    1;2;
2  200    2;3;

Wanted output:
Invoice_ID Names     Amount:
1          Alex;Jim; 100
2          Jim;Tom;  200

我知道数据库模型在技术上并不合适，但现在我使用REGEXP_LIKE的查询来在listagg（）中添加名称，但性能非常慢。

SELECT inv.id as invoice_id, 
       SELECT listagg(n.name,';') WITHIN GROUP (ORDER BY NULL) from names where REGEXP_LIKE(inv.names, '(^|\W)' || n.name || '(\W|$) 'as names
       inv.amount as amount
from invoices inv;

您能否告诉我如何改进此查询？

谢谢！

Answer 1

如果您的names表上有索引，那么这应该更快：

select i.id,
       i.amount,
       (select listagg(n.name, ';') within group(order by null)
          from (select trim(regexp_substr(str, '[^;]+', 1, level)) ASPLIT
                  from (select i.name_ids as str from dual)
                connect by regexp_substr(str, '[^;]+', 1, level) is not null) t
          join names n
            on n.id = t.asplit)
  from invoices i;

即。首先按Name_IDs拆分并加入names表。

Answer 2

我建议您将数据模型更改为以下内容：

NAMES:
ID_NAMES NAME
10       Alex
20       Jim
30       Tom

INVOICES:
ID_INVOICE Amount
1          100
2          200

INVOICE_NAMES:
ID_INVOICE  ID_NAME
1           10
1           20
2           20
2           30

祝你好运。

替代oracle中的REGEXP_LIKE

2 个答案: