[OpenSER-Devel] db_berkeley : METADATA_DEFAULTS

[Will Quan]

Henning,
Thanks for implementing that part of the script!
It should work just like you have it now, but I'll need to test it out.
The driver does not make a distinction between 'NULL' | 'NIL' | 'NO'.
Thus, we can stick with your convention.
Again, Thanks alot.
--Will


Henning Westerholt wrote:
> On Tuesday 13 November 2007, Will Quan wrote:
>> Henning,
>> Sorry to be vague. Lets consider table 'address', as an example. I
>> selected it only because it has 3 columns that have default values.
>>
>> It has a total of 5 columns, like this:
>>    id(int) grp(int) ip_addr(str) mask(int) port(int)
>>
>> The default values for this table are:
>>    grp -> 0
>>    mask -> 32
>>    port -> 0
>>
>> The remaining columns do not specify a default value, so we fill in
>> 'NULL' for those.
>>
>> Thus, the *new* row we would like create would look like this:
>>
>> METADATA_DEFAULTS
>> NULL|0|NULL|32|0
>>
>> So I think the XSL logic would be to always use the default if
>> specified; else use 'NULL'.
> 
> Will,
> 
> i finished the implementation of this script, but 've some more questions.
> This is for example the row for the location table:
> 
> METADATA_DEFAULTS
> NO|''|''|''|NULL|NULL|'2020-05-28 21:32:15'|1.0|'Default-Call-ID'|13|'1900-01-01 00:00:01'|0|0|''|NULL|NULL
> 
> Is the handling of the string values correct for you? I didn't use the 'NULL' value
> for not specified values, as there already exists a NULL type for SQL NULL.
> 
> Should i go for the 'NO' tag, or perhaps use 'NIL'? Do you make a difference for
> not definied values (the 'NO' in the row above), and values specified with SQL
> NULL in the db_berkeley driver?
> 
> Cheers,
> 
> Henning