Re: Pipelined table functions can be used inside a package?

Richard Kuhler <noone_at_nowhere.com> wrote in message news:<6sp4c.24487$Zo6.1016_at_twister.socal.rr.com>...
> There are many bugs related to pipeline table functions still. I've
> filed at least 5 myself and I found another one last week. The one you
> are hitting is likely Bug No. 2982476.

Where can I find the description of these bugs?

I moved all pipelined table functions outside the package and everything works but I don't like this solution. Having all the code in one package is a smartest solution and is very strange that Oracle don't support this.

I'm committed to build a solution that extracts data from the database, transform this data and write the results in a text file. I think that a Java application that connects to the database, make a query like:
SELECT field1, field2, ... fieldN FROM TABLE(my_function(parameters)); and formats the resultset in a text file can be a good solution. The advantage of this solution is that the formatting job (java app) and the extraction job (pipelined table function) can be made in parallel.

I like this solution for many other reason (clear code and clear data flow, possibility to reuse base transormation for more complex transformations, etc... ) but now that I know thare are many bugs on pipelined table functions I'm in dubt to use it. I'm also serching for the optimal performance solution.

Can someone tell me if I'm on the right or on the wrong way?

Thanks and excuse me for my bad english. Received on Sun Mar 14 2004 - 05:36:47 CST