#DateForumTypeThreadPost
52826 May 2009 @ 11:07 PDTDatabaseReplyPDEPDE (Parallel Database Extensions) runs on top of the OS, it's a kind of emulation of the old V1 Teradata Operating System (TOS).Thus PDE is the foundation of Teradata, it may be started without st...
52726 May 2009 @ 10:54 PDTDatabaseReplyDATABASE BACKUPCheck the "Teradata Archive/Recovery Utility Reference" manual.If it's a small system like TD Express, it might be a simple ARC script submitted on command line:LOGON systemA,username,password;ARCH...
52626 May 2009 @ 10:49 PDTAnalyticsReplyTeradata ServicesHi Raj,what's your problem?Don't you want to restart?Did you restart and something is still failing?Check windows Eventlog->Application for Teradata entries.Dieter
52526 May 2009 @ 10:46 PDTAnalyticsReplydifference between distinct and group byIt depends.as a rule of thumb: If you don't know exactly about the data, better use GROUP BY.DISTINCT is better for columns with a low number of rows per value: #rows < #AMPsGROUP BY is better f...
52426 May 2009 @ 10:39 PDTDatabaseReplyHow to view the Teradata temp table dataHi Raju,it's a nice trick for a magician :-)if you do a SHOW TABLE you'll see a "ON COMMIT DELETE ROWS", this is the default postulated by Standard SQL.Your session obviously runs in Teradata mode ...
52322 May 2009 @ 03:38 PDTDatabaseReplyCHAR-COMPRESS vs VARCHAR datatypeHi Koushik,COMPRESS is only usefull for large tables, i wouldn't call 4 rows "large" :-)For each table there's a table header stored on each AMP, check the size of the empty table.After inserting 4...
52222 May 2009 @ 03:23 PDTUDAReplyPerformance issue with the queryAhem, Teradata is not Oracle and posting a PL/SQL script over here is quite useless :-)18m rows should be a piece of cake, unless there's a bad PI on a SET table, as Fred already pointed out.Dieter
52112 May 2009 @ 04:29 PDTDatabaseReplyRows updated insertedHi Monika,there's a Preprocessor/Embedded SQL available for Teradata, but i never worked with it.I'd prefer to write it as a SP, this is the easiest way.Dieter
52012 May 2009 @ 04:22 PDTDatabaseReplyWhat's the quickest way to find out all the values under an attributesThe BETWEEN condition is just wrong and useless:wrong: the first value must be <= the second value and your's vice versa.useless: you need to use DATE or CURRENT_DATE instead of CRTN_DTTM, else ...
51905 May 2009 @ 01:37 PDTDatabaseReplyFETCH NEXT statementCould you please post the actual code?Looks like a missing semicolon...Btw, cursors used for inserts are really bad, this is not Oracle, better rewrite it as an SQL insert/select.Dieter
51804 May 2009 @ 12:15 PDTDatabaseReplyRows updated insertedHi Monika,ACTIVITYCOUNT (like ERRORCODE & ERRORLEVEL) in BTEQ is only available with .IF, but you can't use it for any SQL statement.Whereas the ACTIVITY_COUNT varaible in a SP is set after every S...
51704 May 2009 @ 12:11 PDTToolsReplyfastload performanceI would expect that 1.5GB to be FastLoaded in less than a minute on a single node.Dieter
51604 May 2009 @ 02:09 PDTToolsReplyfastload performanceIt's probably due to the PC the demo is running on, it's using two flat files instead of raw devices.So the bottleneck is you disk, expecially if you load 1.5GB in a 4GB system.Dieter
51504 May 2009 @ 02:06 PDTDatabaseReplyRows updated insertedA NOTIFY exit is written in C, there are examples in the BTEQ manual and the directory where BTEQ is installed.Within that C programm you can do whatever you want, including writing to a flat file ...
51404 May 2009 @ 02:04 PDTDatabaseReplyPE and AMP setupA PE doesn't "serve" AMPs, it's "serving" sessions.Thus there's no ratio between PEs and AMPs, typically you got 2 PEs per node and enough AMPs to keep your hardware (CPU & disks) busy. Dieter
51304 May 2009 @ 02:01 PDTDatabaseReplyWhich space occupies stored proceduresSPs in Teradata are like Embedded SQL programs compiled by a C-compiler.The created .so or .dll file is stored as a (fallbacked) row in the database, thus two AMPs need permspace.Dieter
51204 May 2009 @ 01:59 PDTDatabaseReplyOver partition with order by in teradata, how to usedIf there's ORDER BY in Oracle (and Standard SQL) the default is a Cumulative Sum, but in Teradata it defaults to a Group Sum.You just have to add ROWS:sum(sum(total_revenue)) over(partition by area...
51101 May 2009 @ 02:08 PDTDatabaseReplyRecursion in a stored procedureHi Jesse,i don't know how to get that within the view, but you can calculate it within the select:WITH RECURSIVE temp_hierarchy(parent_id, id, hier_name, lvl) AS (SELECT parent_id, id, hier_name, 1...
51001 May 2009 @ 09:16 PDTDatabaseReplyRows updated insertedHi Monika,there are several possible solutions:1. Parse the output for " *** Query completet." using grep and use awk to extract the count2. Write a NOTIFY exit and use it before each DML statement...
50901 May 2009 @ 09:06 PDTDatabaseReplyRecursion in a stored procedureNo, it works for any tree-like hierarchy and returns *all* children of parent_id x.Dieter
50801 May 2009 @ 05:28 PDTToolsReplyMultiple IMPORT in a LOAD - sequencial or parallel?Hi Akshay,are you talking about MLOAD?I didn't know about that DML limit (and i hope i'll never have to write/maintain a load script where that limit is hit).For your concerns i'd suggest start rea...
50701 May 2009 @ 05:00 PDTDatabaseReplyHow to Learn TeradataHi kishore,Docs\Getting_Started\Getting_Started.htmand for the documentation there's Docs\Teradata_Documentation\Start_Here.pdfHere it's best to start with Teradata Database 12.0-General Reference-...
50601 May 2009 @ 04:48 PDTDatabaseReplyRecursion in a stored procedureHi Jesse,your query just creates the hierarchy tree from root, whereas my query creates the Transitive Closure of that hierarchy.It starts with all rows and then creates all possible pathes up to t...
50525 Apr 2009 @ 04:06 PDTDatabaseReplyRecursion in a stored procedureHi Jesse,you probably did something wrong during search & replace.Just try the query without view:with RECURSIVE Database_Hierarchy (OwnerName, DatabaseName, lvl) AS(SELECTOwnerName,DatabaseName,1 ...
50424 Apr 2009 @ 01:55 PDTDatabaseReplyRecursion in a stored procedureHi Jesse,all children of dbc.Dieter

Pages