Here we go ...
1. Create a logical directory (connected as SYSTEM) in Oracle using ...
CREATE DIRECTORY <DIRNAME> AS '<OSPATH>';
(don't include a '/' or a '\' at the end of the OS path)
2. Grant read access to the directory to the 'load user' ...
GRANT READ ON DIRECTORY <DIRNAME> TO...
I have a table ...
ID TXT
1 dogfood
1 catfood
1 catfood
1 catfood
2 dogfood
3 dogfood
3 <null>
4 <null>
5 dogfood
5 dogfood
5 dogfood
and I want it to produce ...
ID TXT
1 mixed
2 dogfood
3 mixed
4 <null>
5 dogfood
(in other words, if all txt values are the same for an id...
Something like the following should work ...
select <mydata>
from
(
select <fulldataselection>
from <srctable>
order by <sortclause>
) subq
where subq.rownum <=20;
This obviously takes a performance hit on large tables, but anything less than 5,000 records should be quick...
Don't know what version of Oracle you're using and I'm by no means a Rollback Segment 'guru', but you could try setting the OPTIMAL parameter which causes automatic shrinking (when I don't know!), or you could use 'alter rollback segment <rbs> shrink to 10M (for example)'.
A useful view is...
You're most welcome. I need to point out that the use of v$open_cursors means that some uncommitted transactions may have been paged out of memory - nothing's perfect in this world. As an afterthought you may find the following useful to show current locks on the database (it can take some time...
Try this ... (require DBA privileges)
column username format a20
column logon_time format a11
column load_time format a20
set recsep off
set lines 140
break on username on logon_time on sid on load_time
select substr(nvl(s.username,'Background'),1,20) username...
If you explicitly define storage for your CLOB columns, make sure that initial_extent is the same as next_extent (I have pctincrease of 0), otherwise you'll find a HUGE index for the CLOB will be created. Don't know if this is addressed in 8.1.6.
PS. Anybody got any tips on improving the...
Oracle used to licence 'Trusted Oracle' separately which I seem to recall added extensions to the RDBMS to provide row level security. May be a lower long term overhead in terms of maintenance if this is still an active product.<br>
<br>
Mark.
Only tried with Oracle 7.3.4 and SQL Server 7.0. I'm sure it'll work with Oracle 8 as well, since all the translation is done through ODBC. <br>
<br>
Mark.
...select statements must be entirely in uppercase for this to work, so, to select current users (assuming synonym is set up) ...<br>
<br>
SELECT * FROM ORACLELINKEDDB..ORACLEUSER.V$SESSION<br>
<br>
should work. If you get an error message referring to incorrect property settings, the remote...
I got off my lazy backside and sorted this out myself.<br>
<br>
a) I used sp_configure to compare the run_value of both databases and found that 'remote query timeout (s)' was different - it ought to have been zero on the server but was in fact 10.<br>
<br>
b) Solution :<br>
<br>
sp_configure...
"One or more properties could not be set on the query for OLE DB provider 'MSDASQL'. The provider could not support a required property."<br>
<br>
Connecting to a linked server (Oracle) from a network install of SQL Server and performing any SQL query gives the above message. When I...
I think all you need to do is put the assignment settings for purch_date, market, and cost inside the loop - it looks like they're set only at the beginning of the procedure.
I need to clarify with an example ...<br>
Reference table structure is (RefID int,RefDS varchar(50)), Data table structure is (DocID int,RefID int,DocDS varchar(50)). Reference table contains :<br>
RefID RefDS<br>
1 Fruit<br>
2 Vegetable<br>
3 Animal<br>
Data table contains...
My select statement to generate the DBGrid in VB6 contains a join from the data table to the reference table so descriptive text is displayed instead of a code number. When I change the text entry (via a dropdown from the appropriate column - button property set to true) the reference table is...
Thanks for the tip - it pointed me in a direction where I realised my SQL statement was badly flawed (grouping on a dataset that was already grouped). Je suis un buffoon!
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.