#DateForumTypeThreadPost
4214 Nov 2007 @ 04:01 PSTDatabaseReplyDistinct V/s Group ByPlease analyse the Explain Plan of both the queries to get more information.I feel the Table could be heavily skewed.Regards,Annal T
4114 Nov 2007 @ 03:12 PSTDatabaseReplyHow do I create an indexFor Unique SI-- Create Unique Index IndexName (Col1,Col2,...Coln) on TablenameFor Non Unique SI-- Create Index IndexName (Col1,Col2,...Coln) on TablenameRegards,Annal T
4014 Nov 2007 @ 03:07 PSTDatabaseReplyHow do I create an indexWhen you say you have pushed the data to TD i assume it was moved to a TD table.Every table would have a Primary index by default.(Either user defined or automatically assigned)See if the PI column...
3912 Nov 2007 @ 03:11 PSTDatabaseReplyExplain planDocuments are available at www.Info.teradata.com.Information on Explain is available in several documents.You can start with "SQL Reference - Statement and Transaction Processing" PDfChapter Name :...
3809 Nov 2007 @ 06:26 PSTDatabaseReplyExplain planStart with simple queries and try to understand their plans and then play with the query and see the impact on the explain plan. Gradually increase the complexity of the queries and analyse.Regards...
3709 Nov 2007 @ 06:23 PSTDatabaseReplyExplain planHi,There are no quick steps available to understand the explain plan completely.Please go through the documentation for a day or two.I've provided some points which can be a start up for you (Hope ...
3606 Nov 2007 @ 11:31 PSTDatabaseReplySoft Referential IntegritySoft RI is just an indication that there is a PK-FK relation between the columns and is not implemented at TD side.But having it would help in cases like Join processing etc.It is expected that thi...
3505 Nov 2007 @ 11:07 PSTDatabaseReplyHow to force duplication of the rows from a small table on all AMPsCollecting stats should help in this case.Generally collecting stats on the small table would help the optimiser realise that the table is actually small and might make it go for duplication.Make s...
3405 Nov 2007 @ 10:56 PSTTrainingReplyVersion 12 certificationTeradata has mentioned that they would start rolling out tests for the next version from Third quarter of 2008. V2R5 exams will also be available in parallel for some time.Regards,Annal T
3304 Nov 2007 @ 12:24 PSTDatabaseReplyHow to force duplication of the rows from a small table on all AMPsYes. Doing any manipulations on the index would prevent it from being used.Dieter,Will preventing an index usage result in duplication ?Is there any relation between index usage and join strategy(D...
3202 Nov 2007 @ 10:52 PDTDatabaseReplyDENSE_RANKAs dieter mentioned its not available in Teradata.Dense_Rank is an ANSI function available in other databases like Oracle.Regards,Annal T
3102 Nov 2007 @ 05:35 PDTDatabaseReplyDENSE_RANKThe other options are really interesting ways to get Dense_Rank :)Regards,Annal T
3002 Nov 2007 @ 05:31 PDTDatabaseReplyDENSE_RANKI'm not sure if the first option would workSELECTe.department_number,dt.rnk AS "DENSE_RANK",e.last_name,e.salary_amountFROM cs_views.employee eJOIN(SELECTsalary_amount,RANK() OVER (ORDER BY salary_...
2902 Nov 2007 @ 04:21 PDTDatabaseReplyDENSE_RANKDense Rank doesnt leave gaps between ranks.For example in normal Ranking operation if two people have same marks, they both would be given the same Rank, say 1.Now the next person would be given Ra...
2802 Nov 2007 @ 03:58 PDTDatabaseReplyAccessing Partition ID for a TableYes. Partition exists from V2R5.But i thought usage of Partition column for purposes like statistics collection etcwas offered to the users only from V2R6.Regards,Annal T
2731 Oct 2007 @ 02:38 PDTDatabaseReplyAccessing Partition ID for a TableWhats the Teradata version you are using?I think the system derived column "Partition" is available only from V2R6.1Regards,Annal T
2630 Oct 2007 @ 07:36 PDTDatabaseReplyColumn compression and Table header sizeHi Dieter,Any specific reason for selecting Min(currentperm) and not Max(Currentperm) for that table?Regards,Annal T
2530 Oct 2007 @ 06:51 PDTDatabaseReplyWhat is Typical rows per value?Its done in the same way.In the count(*) option select the maximum one.In the above example you can see that the max rows per value is 5.Here the typical rows per value and max rows per value dont ...
2430 Oct 2007 @ 06:00 PDTDatabaseTopicCluster level Archive!!!Hi,I would like to know about "Cluster Archive" and how it differs from an All-Amp Archive.Regards,Annal T
2330 Oct 2007 @ 02:23 PDTDatabaseReplyPPII think joes explanation was complete and clear.With USI, you hash the usi value and go to the subtable which gives exact location of that record which contains this USI value.I assume the subtable...
2230 Oct 2007 @ 01:54 PDTDatabaseReplyColumn compression and Table header sizeI remember seeing in some document that the max size of the compressed values cant exceed 8192 characters.So make sure you dont compress too many values(across columns) which exceeds this size.Rega...
2130 Oct 2007 @ 01:47 PDTDatabaseReplyWhat is Typical rows per value?You can use this value along with "Max rows per value" to see if there are chances of data skew.If those two value differs by a big margin then there are chances of high data skew which might lead ...
2024 Oct 2007 @ 04:05 PDTDatabaseReplyName convension problem.If it says object doesnt exist, try to create a table with the same name again and see what happens ;)Regards,Annal T
1923 Oct 2007 @ 03:03 PDTDatabaseTopicDoubt on the Index Validation part of the Index Wizard!!I worked on the Index wizard available in the Demo package and I would like to know how we use the validate index option in real time.It was stated that since we often use the index wizard in the T...
1823 Oct 2007 @ 02:40 PDTDatabaseTopicScheduling policies in Priority Scheduler !!!I would like to know the difference between "Default" and "Immediate" scheudling policies available in Priority Scheduler.Both are used when we dont want a hard ceiling on the resource usage.It was...

Pages