问题

I have a Column in a Database which contains multiple Values in one Column, which i need as different rows. The Column contains comma delimited parts but also a Part with comma in brackets. I don't need to split this parts. (Only split on commas which are NOT in brackets)

Versions

Oracle 11g

Example:

**ID | Kategory**

1 | "ATD 5(2830),ATO 4(510),EDI 1,EH A1,SCI 2,SS 1,STO-SE 1(oral, CNS, blood),STO-SE 2(oral, respiratory effects)"

This string i need as

 - 1 => ATD 5(2830)
 - 1 => ATO 4(510)
 - 1 => EDI 1
 - 1 => EH A1
 - 1 => SCI 2
 - 1 => SS 1
 - 1 => STO-SE 1(oral,CNS, blood)
 - 1 => STO-SE 2(oral, respiratory effects)

Parts like (oral, CNS, blood) which contains comma in brackets i don't need to split.

回答1:

You can use the regular expression (([^(]*?($.*?$)?)*)(,|$) to match:

[^(]*? Zero-or-more (but as few as possible) non-opening-bracket characters
($.*?$)? Then, optionally, an opening bracket and as few characters as possible until the closing bracket.
( )* Wrapped in a capturing group repeated zero-or-more times
( ) Wrapped in a capturing group to be able to reference the entire matched item
(,|$) Followed by either a comma or the end-of-string.

Like this:

SQL Fiddle

Oracle 11g R2 Schema Setup:

CREATE TABLE table_name ( ID, Kategory ) AS
SELECT 1, 'ATD 5(2830),ATO 4(510),EDI 1,EH A1,SCI 2,SS 1,STO-SE 1(oral, CNS, blood),STO-SE 2(oral, respiratory effects)' FROM DUAL;

Query 1:

SELECT ID,
       l.COLUMN_VALUE AS item,
       REGEXP_SUBSTR(
         Kategory,
         '(([^(]*?(\(.*?\))?)*)(,|$)',
         1,
         l.COLUMN_VALUE,
         NULL,
         1
       ) AS value
FROM   table_name t
       CROSS JOIN
       TABLE(
         CAST(
           MULTISET(
             SELECT LEVEL
             FROM   DUAL
             CONNECT BY LEVEL < REGEXP_COUNT( t.Kategory, '(([^(]*?(\(.*?\))?)*)(,|$)' )
           )
           AS SYS.ODCINUMBERLIST
         )
       ) l

Results:

| ID | ITEM |                               VALUE |
|----|------|-------------------------------------|
|  1 |    1 |                         ATD 5(2830) |
|  1 |    2 |                          ATO 4(510) |
|  1 |    3 |                               EDI 1 |
|  1 |    4 |                               EH A1 |
|  1 |    5 |                               SCI 2 |
|  1 |    6 |                                SS 1 |
|  1 |    7 |          STO-SE 1(oral, CNS, blood) |
|  1 |    8 | STO-SE 2(oral, respiratory effects) |

来源：https://stackoverflow.com/questions/48376315/split-columntext-to-rows-extract-delimiter-in-bracket-oracle-sql

标签

Oracle