I don't see how the problem can be solved by merging the function logic into the calling function. The int[][] array is
suppliedby user, and I still have the issue to pass it into a dynamic statement, no matter in which function, unless I
serializeit to string and deserialize back to int[][].
----- Original Message ----
From: Robert Haas <robertmhaas@gmail.com>
To: Sophie Yang <yangsophie@yahoo.com>
Cc: Pavel Stehule <pavel.stehule@gmail.com>; pgsql-hackers@postgresql.org
Sent: Thursday, March 5, 2009 6:51:48 PM
Subject: Re: [HACKERS] Use array in a dynamic statement
On Thu, Mar 5, 2009 at 8:32 PM, Sophie Yang <yangsophie@yahoo.com> wrote:
>
> I was wondering why USING clause is not supported in pl/pgsql dynamic statement. Serialization is the option I tried
toavoid, but it seems there is no better approach available.
>
> Just to say a few more about the usage of my function. In dag_tree_1, (rid, rtid) is the primary key, which
identifiesa node in a tree structure. The idx field is a kind of dewy index. for example:
> rid rtid idx
> 1123 1 0000.0006.0033
> 3231 1 0000.0006
> 786 6 0000.0007.8853
> 80923 2 0000.0007.8853.2382
>
> The function takes in a list of rid and rtid pair (nids), sort them by the length of the dewy index, which is
equivalentto sort the nodes by their tree depth. That's what I try to achieve. Maybe someone has different idea to
implementthe function?
Well the function is pretty simple. Maybe you could merge the logic
encapsulated by the function into the containing query?
...Robert