#DateForumTypeThreadPost
55314 Jul 2009 @ 08:25 PDTUDAReplyUsing replace in TDVarchar(500) should work, check the definition/source of the oreplace function: * NOTE: The define MAXIMUM_LENGTH must match the VARCHAR length for * the return parameter in the CREATE/REPLACE FUNC...
55214 Jul 2009 @ 08:16 PDTDatabaseReplyDBQLOGTBL SPOOL USAGE NOT CORRECT IN TD12Is the PMon values less or greater than DBQL?Are you shure you presses the refresh button in PMon the right time?Could you post some numbers?If this repeats for certain types of queries you should ...
55114 Jul 2009 @ 08:09 PDTDatabaseReplyAccess ProblemHi Ansh,but you still have to submit GRANT CREATE TABLE ON dbname TO dbnamr WITH GRANT OPTIONDieter
55014 Jul 2009 @ 07:48 PDTDatabaseReplyUsing Insert Mode in TeradataHi Ansh,an insert/select into an empty table might be a so-called "fast-path" insert/select, if the target table is empty at the beginning of the transaction.If you run tose statements in Teradata ...
54914 Jul 2009 @ 07:40 PDTDatabaseReplyAccess ProblemHi Ansh,the database where the macro is located needs that access right.Dieter
54814 Jul 2009 @ 07:38 PDTUDAReplyFormatting decimal fields while exportingHi Chaitali,did you work with Oracle before? :-)AFAIK there's no format for suppressing trailing zeros.TRIM(BOTH '0' FROM col (FORMAT '9(3).9(4)')) || CASE WHEN col MOD 1 = 0 THEN '0' ELSE '' END ...
54714 Jul 2009 @ 07:02 PDTToolsReplyImproving Loading Mechanism using FastloadTPT uses the same load protocols as FastLoad or MLoad, if TPT will be faster depends where's the bottleneck. Could you post a log output again?Dieter
54613 Jul 2009 @ 03:23 PDTToolsReplyACCOUNT STRING FOR TPT OPERATORSHi Ron,of course, but it's the 4th parameter, because the first is the optional tdpid :-)Dieter
54513 Jul 2009 @ 03:10 PDTUDAReplyUsing qualify and group by togetherHi Hrushikesh,OLAP functions are processed after GROUP BY/HAVING, you probably try to access a column which is neither in GROUP BY nor aggregated.Could you post your actual query?Dieter
54413 Jul 2009 @ 02:03 PDTDatabaseReplyUsing Insert Mode in TeradataHi Ansh,#1 is one multi-statement request, if it's submitted using BTEQ, whereas #2 are two request.The number of rows inserted will be the same (unless there's an error).Check the manuals for deta...
54313 Jul 2009 @ 01:58 PDTDatabaseReplyStrange ProblemHi Ansh,run a HELP SESSION; and check TRANSACTION SEMANTICS, if it's ANSI or TERADATA.It's probably ANSI, so you have to submit a COMMIT; before you logoff your session.Dieter
54213 Jul 2009 @ 01:55 PDTDatabaseReplyqualify rank() over (partition.....questionYou do exactly the same, you would do in Oracle :-)Replace RANK with ROW_NUMBER:qualify row_number() over (partition by cust_acct_nbr order by cnt desc, commodity_desc) = 1Dieter
54105 Jul 2009 @ 10:35 PDTDatabaseReplySelect clause inside fields listHi Rodrigo.you can rewrite it using an OLAP function, too:SELECT ID_CDR, COUNT(*) AS QTD ,CAST (QTD * 100 / sum(QTD) OVER () AS DEC(10, 2)) AS PercentFROM Table_1 GROUP BY ID_CDRQUALIFY ROW_NUMBER...
54002 Jul 2009 @ 01:52 PDTDatabaseReplyDBQLOGTBL SPOOL USAGE NOT CORRECT IN TD12And comparing it to what?Dieter
53902 Jul 2009 @ 01:52 PDTDatabaseReplyInsufficient memory when I perform a commandHi Anderson,you're submitting a huge multi-statement request, too huge to parse it as a single block.- split it into several smaller batches- Or better check if iterated request are available for ....
53802 Jul 2009 @ 01:44 PDTDatabaseReplydrop journal tables.How did you create the journal? This is how to drop it, too.modify database xxx as drop default journal table;Dieter
53702 Jul 2009 @ 01:39 PDTAnalyticsReplyAnother Question regarding DBC.AMPUSAGEThe account string is expanded once when a query starts. So it's just one row with the start date/hour and accumulated cpu/io.Dieter
53602 Jul 2009 @ 02:57 PDTAnalyticsReplydbc.ampusageDepending on "account string expansion" (ASE) you might have more than a single row per user, e.g. one row per day or hour or request.Before there was the QueryLog this was the only way to get more...
53502 Jul 2009 @ 02:40 PDTToolsReplyFastLoad Error -incorrect number of bytes returnedHi Manish,are you running that script on a Teradata Express, i.e. Windows?Then you have to change the newlinechar to a char(2), because on Unix it's just line feed '0A'xc whereas on Windows it's ca...
53430 Jun 2009 @ 10:13 PDTAnalyticsReplycustomer segmentationHi Vas,could you post some more details, especially DDL & some INSERTs & expected answer set?This sounds like a task for OLAP functions.Dieter
53330 Jun 2009 @ 09:52 PDTAnalyticsReplyLevenstein Distance AlgorithmI don't know if this is what you actually need, but there's a UDF implementing an editdistance funtion athttp://www.teradata.com/DownloadCenter/Topic9232-133-1.aspxDieter
53230 Jun 2009 @ 09:37 PDTDatabaseReplyMeasuring Query reponse time.The Full Table Scan is finished after 1.5 seconds and all data is in spool now:Teradata starts sending back the first row *after* the whole answer set has been created, so for the PE (which maintai...
53130 Jun 2009 @ 07:07 PDTToolsReplyPE and AMPYou're correct, in Teradata Version 1 every AMP/PE was a single proprietary server with Memory, CPU, disks and Operating System. In V2 it was rewritten as software only, now those processors are on...
53030 Jun 2009 @ 07:01 PDTDatabaseReplyMeasuring Query reponse time.Forget about LastResponseTime, it never actually worked, that's why it's removed in TD12.After ElapsedTime the first rows are send back to the client and this is when users start to look at the res...
52926 May 2009 @ 11:36 PDTDatabaseReplyGMT to EDT conversion in Teradatainventory.gmt_start_time is a string with a GMT=UTC timestamp and must be converted to EDT (= -4 hours)?And minutes should be truncated to half hours?string -> timestamp: cast(strcol as timestam...

Pages