REGEXP_COUNT ignorerer store og små bogstaver

Kør:

SELECT *
FROM   NLS_SESSION_PARAMETERS
WHERE  parameter IN ( 'NLS_COMP', 'NLS_SORT' );

Og se, hvilke NLS-sessionsparametre du bruger.

For eksempel:

ALTER SESSION SET NLS_COMP=BINARY;
ALTER SESSION SET NLS_SORT=BINARY;
SELECT regexp_count('OOO', '[a-z0-9]', 1, 'c' ) AS COUNT_AZ,
       regexp_count('OOO', '[[:lower:][:digit:]]', 1, 'c' ) AS count_lower
FROM   DUAL;

Udgange:

Men:

ALTER SESSION SET NLS_COMP=ANSI;
ALTER SESSION SET NLS_SORT=GERMAN_AI;
-- or ALTER SESSION SET NLS_SORT=GERMAN_CI;
-- or ALTER SESSION SET NLS_SORT=FRENCH_CI;
-- or many others
SELECT regexp_count('OOO', '[a-z0-9]', 1, 'c' ) AS COUNT_AZ,
       regexp_count('OOO', '[[:lower:][:digit:]]', 1, 'c' ) AS count_lower
FROM   DUAL;

Udgange:

For at rette dette kan du enten:

• Skift NLS_SORT sessionsparameter.
• Eller (som nævnt af @AlexPoole) brug [:lower:] POSIX-tegnsæt i stedet for a-z .

db<>fiddle her