RE: workaround for odciaggregate bug?

From: Stauffer, Robert G <>
Date: Tue, 27 Jan 2009 14:27:04 -0500
Message-ID: <>


Thanks. Your query does the trick.

And, true, it doesn't make sense to expect the row ordering to work - especially since ODCIAggregate seems to have been targeted at numeric aggregation.

Thanks to Ian, Mark, and Charles for your help, too.

Bob Stauffer
D&E Communications
Ephrata, PA
-----Original Message-----
From: Stephane Faroult [] Sent: Friday, January 23, 2009 17:26
To: Stauffer, Robert G
Subject: Re: workaround for odciaggregate bug?


    I'm not sure it really qualifies as a bug, because row ordering is something that has no meaning in the relational theory, and therefore Oracle is free to return rows in whichever order it wants - only the final "order by" counts.
But there is another solution than stringAgg, and this one works "by design":

SQL> select acctnbr,
  2 replace(path, '|', '') result   3 from (select acctnbr,

  4            sys_connect_by_path(prodcount, '|') path,
  5            rn,
  6            cnt
  7         from (select acctnbr,
  8              prodcount,
  9              row_number() over (partition by acctnbr
 10                         order by snapshotdate) rn,
 11              count(snapshotdate) over (partition by acctnbr) cnt
 12           from testtable)
 13         connect by acctnbr = prior acctnbr
 14            and rn = 1 + prior rn
 15         start with rn = 1)

 16 where rn = cnt
 17 order by acctnbr
 18 /


---------- --------------------

0123456789 010203

HTH S. Faroult

Stauffer, Robert G wrote:
> All,
> and on AIX 5.2
> I'm trying to use ODCIAggregate and Tom Kyte's stringAgg function to
> concatenate a number in multiple rows into a string. Below is the
> output of the select statements I ran. Create scripts for the table,
> type, and function are below the output. I'm expecting the returned
> string to be '010203' but it returns '010032'. Unfortunately, it
> like there's a bug (7194959 to be fixed in 11.2) in ODCIAggregate that
> "unsorts" the sorted data when it creates the concatenated string.
> looked around (asktom, metalink, etc.) and can't seem to find a
> work-around for this method. Has anyone come up with one?

This e-mail message and any files transmitted with it are intended for the use of the individual or entity to which they are addressed and may contain information that is privileged, proprietary and confidential. If you are not the intended recipient, you may not use, copy or disclose to anyone the message or any information contained in the message. If you have received this communication in error, please notify the sender and delete this e-mail message. The contents do not represent the opinion of D&E except to the extent that it relates to their official business.

Received on Tue Jan 27 2009 - 13:27:04 CST

Original text of this message