Regular Expression TO validate column name

Question

Hi All
I want to write a regular expression to validate column names.i.e enterted name should follow the valid column name convention
contain only A–Z, a–z, 0–9, _, $, and #
Please someone help me to do this.
Padma

Answer 1

Hi,
I would write it that way :SQL> with t as (
  2          select 'ABC$DEF1' cname from dual
  3          union all select 'ABC#DEF2' from dual
  4          union all select 'ABC_DEF3' from dual
  5          union all select 'ABC_#$x4' from dual
  6          union all select 'abc_def5' from dual
  7          union all select 'ABCDEFG%' from dual
  8          union all select 'ABC*DEFG' from dual
  9  )
10  select cname
11  from t
12  where not regexp_like(cname,'^[_#$[:alnum:]]+$') ;
CNAME
ABCDEFG%
ABC*DEFG
2 rows selected.You could also add some length condition in the regexp if your convention is supposed to enforce any.
Just my 2 cents :
DBMS_ASSERT.SIMPLE_SQL_NAME looks good +(I didn't know that package)+ but it would allow double quoted names with "unwanted" characters inside. Even if valid for the RDBMS maybe it's not ok regards to the OP's naming convention :SQL> l
  1  with t as (
  2     select '"Name with blank spaces"' cname from dual
  3     union all select '"Name with @ stuff"' from dual
  4     union all select '"Name with ~ stuff"' from dual
  5  )
  6  select DBMS_ASSERT.SIMPLE_SQL_NAME(cname) checked
  7* from t
SQL> /
CHECKED
"Name with blank spaces"
"Name with @ stuff"
"Name with ~ stuff"
3 rows selected.I guess this procedures lacks some option parameter to specify what is valid or not.
+(I hate to see column names with blank spaces in it)+
(^_^)
More seriously, if the point is to follow a chosen namig convention, I guess it would make more sense to have it explicitly written by the owner.
(You can't say how the DBMS_ASSERT package will evolve in futures RDBMS releases).