Call to_char, to_number or to_date with nls_pars but without a format model

Question

Hi
     Does somebody know if there is a way to call to_char, to_number or to_date functions defining nlsparam argument and without defining the format model arument, in order to use implicit conversions.
Thanks
Bye Alessandro

Answer 1

In my environment there is a table where I store old values (each field as varchar2) of updated and deleted table records with a sort of audit_triggers.
These triggers are the ones using to_char(<number>) to convert a number. It works fine because it does what I need. If i specify a format model ( the same for all the conversion ) I have some problems. Some of them are the following, there are more.
- with decimals for the format model on an integer I'll have a conversion like "1.00000000000" on the table storing old values and I had indications to avoid this behavior.
- if I don't specify enough 9s for the integer part the number will be converted to a string like "################".
Now I found that some users work with a different nls_numeric_characters session parameter and this introduced problems for reading those numbers because 1.500 can be one dot five or fifteen hundreds.
A format model that does the same as implicit conversion would work as well but I didn't find it yet.
Thanks Alessandro
Message was edited by:
Alessandro Rossi
My need then is to call to_char() with nls_pars but without a format model
A sample of why I want the implicit conversion
select to_char(1989793456895722342342334534534564356435902636)
from dual
select to_char(198979345689572234234233453453456435253453154153456435902643536,'999999999999999999999999999999999999999999999999999999')
from dual
/