I have a few Oracle tables that have gotten way too big. Normal
DELETE/WHERE clauses consume too many resources. I know enough
to know that cursors would be more efficient here but I don't
know enough to actually start using them. Does anyone have an
example or two (or three) to bootstrap me?
Hi Folks
Sorry if you've seen this, but I did not receive a copy.
(This is under WinXP, with a DSN pointing to an MS Access db).
Is there an ODBC function call, or anything, which will distinguish between
tables and views?
I looked at DBI::Const::GetInfo::ODBC, and tried $dbh - get_info(19)
The last two arguments to SQLTables can take a string containing the types you
want:
SQLTables(stmt_handle,
NULL, 0, /* no specific catalog */
NULL, 0, /* no specific schema */
NULL, 0, /* no specific table */
NULL, 0) /* no specific type - table
Steve,
maybe this whole thing doesn't really belong here but anyway:
Steve Sapovits wrote:
I have a few Oracle tables that have gotten way too big. Normal
DELETE/WHERE clauses consume too many resources. I know enough
to know that cursors would be more efficient here but I don't
How do
On Wed, Jun 22, 2005 at 05:27:28PM -0700, Jesse Molina wrote:
Greetings
I am having difficulty making and make testing DBD::Oracle.
Attached are files with detailed info which is asked for in the README
file. Hopefully the listserver will accept them.
My Oracle client is 8.0.5. I
Steve,
Delete from table where
will be much much more efficient than anything involving a cursor
and a where current of clause to delete the row you are iterating over the
cursor.
think about what delete does. it identifies only the rows that need to be
deleted, and deletes them in
Daniel,
On 22-Jun-2005 Daniel Kasak wrote:
Martin J. Evans wrote:
As far as I am aware this is a limitation of the TDS protocol used and
your example code will not work unless you force a server-side cursor (see
DBD::ODBC pod and look for odbc_SQL_ROWSET_SIZE but I warn you that you
really
Steve:
I have a PL/SQL process that I use to purge rows from a table
that grows by about 14M rows/month. Even though it's PL/SQL,
the process should be (might be?) transportable to Perl.
Basically, it looks like this:
Define SQL statement to gather the rows (PL/SQL cursor);
ctr = 0;
For each
Well, I have to echo what others have said. The only reason I can think
of for using loops to do your delete is if you really want to save a
chunk of rollback segment. Even then, I would consider batching the
rows up into chunks that you could manage with predicates like BETWEEN
or IN.
dittoand I'd add that if your delete is taking forever to run that
you may want to consider getting some indexes on the columns you are
using to filter for your delete. If a select * from myTable where
insert your predicate here takes a long time, then so will a delete
with the same
Size your rollback so you can do this in one
transaction.
DECLARE
CURSOR T1 IS
SELECT e, f
FROM T1
WHERE e f
FOR UPDATE;
BEGIN
OPEN T1Cursor;
LOOP
/* Retrieve each row of the result of the
above query
into PL/SQL
sorry for the earlier aborted message.
yahoo mail mysteriously sends messages while I am
still typing sometimes. some magic key I hit on my
laptop submits the form.
I was including a sample of where current of
but before that:
1. Size Rollback for non-incremental commits
2. Don't do Manual
Hi Tim
Thanks for your reply. Sorry if I was not clear, but I had already tried your
suggestion and it did not work.
Specifically, in the Makefile.PL, I did;
[EMAIL PROTECTED]diff Makefile.PL Makefile.PL.original
287c287
$opts{LIBS} = [ -L$OH/$libdir -lextp -l$lib $syslibs ];
---
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Version 1.43 of DBD::Pg, the Perl DBI interface to PostgreSQL,
has just been released. This version mostly fixes some bugs
that appeared in 1.41 and 1.42, as well as introducing a new
file, README.dev, to help developers of this module.
The module
check out TOYS (Tool for Organizing Your Schemas)
http://www.impacttoys.com
John Watson
Steve Sapovits [EMAIL PROTECTED] 06/16/05 11:11AM
Does anybody know of any good tools (free or otherwise) for comparing
two Oracle schemas? I'd want to be able to compare tables, indices,
sequences, etc.
On Thu, 23 Jun 2005 11:49:01 -0400, Cees Hek wrote:
Hi Cees
I think that if we want complete control over ordering (ie maximal
I just love this phrase! Nothing personal, Cees, I assure you.
A lot of the problems uncovered in this discussion are due to attempts to use
code to solve the wrong
16 matches
Mail list logo