您可以通过提取不同的捕获组（用()圆括号括起来）来避免双重反转：
WITH t ( VAL ) AS (
SELECT 'my_new_table_2015_06_31' FROM DUAL UNION ALL
SELECT 'my_new_table_temp_2016_06_31' FROM DUAL
)
SELECT REGEXP_SUBSTR( val, '^(.*)_([^_]+)_([^_]+)_([^_]+)$', 1, 1, NULL, 1 ) AS COL4,
REGEXP_SUBSTR( val, '^(.*)_([^_]+)_([^_]+)_([^_]+)$', 1, 1, NULL, 2 ) AS COL3,
REGEXP_SUBSTR( val, '^(.*)_([^_]+)_([^_]+)_([^_]+)$', 1, 1, NULL, 3 ) AS COL2,
REGEXP_SUBSTR( val, '^(.*)_([^_]+)_([^_]+)_([^_]+)$', 1, 1, NULL, 4 ) AS COL1
FROM t
您甚至可以通过使用以下命令使正则表达式简单得多：
'^(.+)_(.+)_(.+)_(.+)$'
第一个.+是贪婪的，因此它将尽可能匹配，直到在第二个至第四个捕获组中只剩下足够的字符串用于最小匹配为止。
：
WITH t ( VAL ) AS (
SELECT 'my_new_table_2015_06_31' FROM DUAL UNION ALL
SELECT 'my_new_table_temp_2016_06_31' FROM DUAL
)
SELECT SUBSTR( val, 1, pos1 - 1 ) AS col4,
SUBSTR( val, pos1 + 1, pos2 - pos1 - 1 ) AS col3,
SUBSTR( val, pos2 + 1, pos3 - pos2 - 1 ) AS col2,
SUBSTR( val, pos3 + 1 ) AS col1
FROM (
SELECT val,
INSTR( val, '_', -1, 1 ) AS pos3,
INSTR( val, '_', -1, 2 ) AS pos2,
INSTR( val, '_', -1, 3 ) AS pos1
FROM t
);
602392714
清零编程群