#DateForumTypeThreadPost
24101 Apr 2016 @ 11:33 PDTDatabaseReplyHow PDE increase Teradata performance?CPUs need to be thought of in terms of cores and hyperthreads, not CPU chips. A CPU chip might be 2Ghz but its total power needs to be considered as the number of cores and hyperthreads in the core...
24031 Mar 2016 @ 07:48 PDTDatabaseReplyfastload to NOPI table gives skewExecuting UpdateSpace and checking again would allow determination whether it is a space accounting issue.    I suggest opening a teradata incident in any case. Skew is expected if the ...
23929 Mar 2016 @ 04:55 PDTGeneralReplyPerformance of Global temp tables vs permanent tables?It depends...   The permanent table will hold perm space until the delete is performed. If the data no longer has value when the session is done, the temp table will use spool and release th...
23829 Mar 2016 @ 04:37 PDTDatabaseReplyEncountered an error using SUM Over Partition byIt sounds like you want the total top up and total revenue for each combination of date, product, region. If that is correct, remove the over... clause completely and use the group by.   ...
23729 Mar 2016 @ 04:16 PDTDatabaseReplyTable size are different from DEV to PRODOr perhaps Dev is an appliance platform with hardware compression built in. 
23629 Mar 2016 @ 04:48 PDTDatabaseReplyPerformance of the We do a MERGE into table HUB_ENCOUNTER_MESSAGE from Spool 18767. step of an insertAn SP debugger doesn't help if it can't be seen in Explain (not that an SP debugger isn't a good idea). I haven't looked recently to see if it can be determined from DBQL?
23528 Mar 2016 @ 03:10 PDTDatabaseReplyPerformance of the We do a MERGE into table HUB_ENCOUNTER_MESSAGE from Spool 18767. step of an insertIn the SP above, there is a single row insert into the table before the INSERT SELECT. That makes the IS a non-empty table IS rather than an empty table one. That means teradata has to TJ all ...
23418 Mar 2016 @ 09:39 PDTPrivate Forum
23324 Feb 2016 @ 05:38 PSTDatabaseReplyFind table size and last access date If an object existed prior to turning on usecount, and it has not been accessed since usecount was turned on, then the last access will still be null.
23223 Feb 2016 @ 10:24 PSTToolsReplyDate value nullTo add a bit of ?why? to Fred's excellent answer...   CAST to character does not result in '?' for NULL values. Any function, arithmetic,... on NULL values results in a NULL valu...
23118 Feb 2016 @ 11:13 PSTDatabaseReplyError compare timesTIME is stored internally at UTC timezone. That suggests that your system or session timezone is not at UTC so a timezone adjustment is being made.  21:00 is likely wrapping around to 00:00 wh...
23010 Feb 2016 @ 01:33 PSTDatabaseReplyTransaction ABORTed due to Teradata table deadlock in InformaticaWhen multiload, fastload or the TPT load/update operator methods are used then multiple sessions can be used without deadlocks. The previous comment can be true if multiple concurrent ODBC ses...
22925 Jan 2016 @ 11:49 PSTDatabaseReplySpool Space Issue on More Projected Columns Than Lesser Projected ColumnsA close look at the explain tells the story. In query 1, the content of Table1 participating in the query is much bigger since there are so many more columns included from it. The optimizer le...
22825 Jan 2016 @ 10:09 PSTDatabaseReplyLong Parsing IssueThere is no ideal parsing time. A very complex query accessing a complex data model can take a while to parse in favor of having a shorter run-time by finding the best plan. Eg if spending a minute...
22721 Jan 2016 @ 12:37 PSTPrivate Forum
22606 Jan 2016 @ 02:08 PSTDatabaseReplyMaterialize (or spool) WITH clause queriesThe query and Explain would be helpful. The optimizer has the capability to automatically reuse the results of any subquery (including WITH) if appropriate.
22524 Dec 2015 @ 07:59 PSTToolsReplyBTEQ - Row Level COMMITWas there another error code/message? Did you Commit? If your session is in ANSI mode then a Commit is required.
22411 Dec 2015 @ 10:27 PSTDatabaseReplyvery fast query + very fast query = very slow query ???It is very likely that there are not statistics on the Description column. 90% is the estimate that the optimizer will use in the absence of statistics. I respect your opinion regarding this defaul...
22311 Dec 2015 @ 09:52 PSTDatabaseReplyvery fast query + very fast query = very slow query ???HELP STATISTICS <database>.<table>; will tell you the current status of statistics on the table.   If there are not statistics on that column: COLLECT STATISTICS COLUMN (<dat...
22211 Dec 2015 @ 08:50 PSTPrivate Forum
22111 Dec 2015 @ 08:21 PSTDatabaseReply3899 Internal error in teradata SQL parserAn Internal error like this should be reported to the Teradata support center. They will most likely need the DDL of the views/tables referenced. They will also likely be able to provide a workarou...
22010 Dec 2015 @ 01:29 PSTDatabaseReplyvery fast query + very fast query = very slow query ???- "Do steps in parallel" in the explain is a different parallelism. It just means that the optimizer has decided that it can execute those steps simultaneously rather than doing them sequ...
21910 Dec 2015 @ 01:17 PSTDatabaseReplyLooking for View Definition in DBC tableSee Dieter's 22 March 2011 note above. .RequestText has what you are looking at.  .CreateText has the complete view definition.
21808 Dec 2015 @ 08:27 PSTDatabaseReplyvery fast query + very fast query = very slow query ???If you look closely at the step that implements edw_tables.sales_master.InventoryDescription LIKE 'Widget B%', you will see an estimate of "with no confidence to be 91,068 rows". ...
21708 Dec 2015 @ 07:32 PSTPrivate Forum

Pages