On Monday, December 10, 2001, at 10:30 AM, Wilson, Doug wrote:

>
>> From: Terrence Brannon [mailto:[EMAIL PROTECTED]]
>>
>> On Monday, December 10, 2001, at 09:16 AM, Jeff Zucker wrote:
>>
>>> Terrence Brannon wrote:
>>>>
>>>> The following SQL is not parseable by SQL::Statement:
>>>>
>>>> INSERT
>>>>    INTO   thot_log (thot_fk,thot_type_fk,thot_temporality_fk,date)
>>>>    VALUES (?,?,?,CURRENT_TIMESTAMP)
>>
>> What I have is a program which reads a SQL statement from a file.
>> It parses it with SQL::Statement, then it prepares this
>> statement, then does an execute against it with @ARGV. I tried
>> your quoting suggestion. In fact, I put the following 3 different
>> things in the file:
>>
>
> Prepares it and executes it against what database? If its 
> Postgres, then
> Postgres should be parsing it, not SQL::Statement. CURRENT_TIMESTAMP is

if you looked at sub SQL::Catalog::test and the associated 
commentary I made, the answer to this would be clear: the 
above-mentioned subroutine parses the statement with 
SQL::Statement so that it can store statistics about the query. 
Then it prepares and executes it via a routine which returns a 
valid $dbh


>  You can't expect SQL::Statement to handle every
> special thing that every database might include in its 
> featureset/syntax.
>

Is Jeff redoing SQL::Statement with Parse::RecDescent? If so, we 
would conceivably  subclass it with SQL::Statement::Postgres and 
I could add the appropriate new grammar items as I needed them.

Reply via email to