#DateForumTypeThreadPost
11611 Jul 2014 @ 12:10 PDTDatabaseReplyAll AMP active for Updates based on primary index- how are you determining that all amps are active for that single update? - are you running multiple updates concurrently - against the same or multiple tables?
11508 Jul 2014 @ 03:13 PDTDatabaseReplyhow can you drop global temp table definition?What error did you get? You are not allowed to drop or modify the GTT definition if there are active instances in any session including your own.   It is definately possible to dro...
11407 Jul 2014 @ 09:42 PDTDatabaseReplyCan we able to perform all the ETL activites by using VIEWS ?????You can do insert, update and delete operations through views as long as you follow all the "updateable view" rules listed in the manual. However, this will not affect performance positiv...
11326 Jun 2014 @ 07:28 PDTDatabaseReplyoptimize the queryIn most of the above joins the tables are related via the .id field. In the first query m is joined to r only on the inequality on the date expression which will force a full product join comp...
11219 Jun 2014 @ 03:40 PDTDatabaseReplyEnd Record in past time as incoming record start date [ TEMPORAL ]Use SEQUENCED UPDATE or SEQUENCED MERGE-INTO. Below is the example   drop table vt1; drop table incoming_vt1; create multiset table vt1( id int, name char(10), valid_dt period(date) as val...
11105 Jun 2014 @ 09:32 PDTDatabaseReplyIdentifying HOTAMP / COLD AMPTotally different question than what started this thread. The original question was about segregating data based on usage frequency. Sorry that the terminology got overloaded and became confusing.....
11003 Jun 2014 @ 09:15 PDTDatabaseReplySQL statements in UNION executes serial or parallelThere is not such an option. It would not work for three reasons under current optimizer/execution engine rules. - Currently the engine is not able to write to the same spool file from multiple co...
10930 May 2014 @ 03:26 PDTDatabaseReplyBest practise table creation with primary key or/and index?PK or UNIQUE in Teradata do not allow the column(s) to be nullable. For the UPI definition, you control whether nulls are allowed by specifying the appropriate null option on the column. If yo...
10830 May 2014 @ 10:22 PDTDatabaseReplyUnique secondary index violation errorThere was a SQL posted that no longer seems to be here but... that SQL used RANK to generate the _op_sys_id column. RANK in Teradata is not guaranteed to generate unique numbers, if two records by ...
10730 May 2014 @ 09:44 PDTDatabaseReplyBest practise table creation with primary key or/and index?A couple comments on the original post: - In Teradata, a UPI on a table will result in only the columns comprising the UPI being checked for uniqueness. Making it a set or multiset table does not ...
10625 May 2014 @ 10:24 PDTDatabaseReplyWhy is this statistic recommended?If you think there can be a better plan, looking at the explain will tell you a lot more than looking at the diagnostic helpstats. Look at the explain and look for places where the estimates are fa...
10517 May 2014 @ 03:49 PDTDatabaseReplyHow to calulate Skew on Join IndexThere is a comment above about NUSIs and skew. This is separate from the JI question being discussed here because a JI can have a completely separate skew from its underlying base table(s) but a NU...
10416 May 2014 @ 03:25 PDTToolsReplyDifference Between Load & Export UtilitiesA few clarifications: (as this is read, all cases of Multiload can be read as TPT Update, Fastload as TPT Insert, TPump as TPT Streams - TPT should be being leveraged at this point for any new deve...
10315 May 2014 @ 12:18 PDTGeneralReplyMload vs Fastload in rejecting the duplicatesPhase 1 - data is supplied from the client/load server side of Fastload/TPT Load Operator in blocks of data sent down multiple sessions each of which is connected to a different AMP (not all AMPs h...
10212 May 2014 @ 04:23 PDTDatabaseReplyTeradata vs oltpI agree with the cautions about locking however running a transaction workload through a single session is very unlikely to deliver the required throughput or respose time unless the volume is extr...
10108 May 2014 @ 09:16 PDTDatabaseReplySkew, amp and column selectionIt never makes a difference. As long at it is the same columns, the hashing will be the same. Nothing you do to the order or the position in the table or the PI will make any difference in the way ...
10025 Apr 2014 @ 01:03 PDTGeneralReplyUtilising Wasted Space in TeradataThe above query is not accurate for a group of tables. It assumes that the skew for each table is on the same AMP which is often/typically not true. Teradata does not allocate equal space to the t...
9918 Apr 2014 @ 10:31 PDTGeneralReplyWhen we need to collect stats on tableThere is a lot of material by Carrie Ballinger here on DevX which provides recommendations on where and when to collect stats. I suggest starting there.
9815 Apr 2014 @ 09:35 PDTDatabaseReplyWhich is efficient?Spool will remain the same, execution cost after the initial table reads will be the same. Overall performance of the query will not be the same because the underlying table has to be read in order...
9710 Apr 2014 @ 08:12 PDTDatabaseReplyNumber of AMPs Needed??As noted by Dieter, there is not a relationship between the amp worker task (AWT) and the session. When a session is logged on but not executing, it uses no AWTs. When a request is submitted, ...
9610 Apr 2014 @ 07:56 PDTDatabaseReplyRounding NumberOr the programmer's trick: cast((col1+.5) AS integer) (cast to integer truncates the decimal portion)
9504 Apr 2014 @ 01:30 PDTDatabaseReplyTeradta Error 6908It appears that parameter arrays are in use. It is not allowed to send a multi-statement request when parameter array is used - that is what this error is saying. The code will need to be reviewed ...
9404 Apr 2014 @ 09:57 PDTPrivate Forum
9302 Apr 2014 @ 01:11 PDTDatabaseReplyVolatile table vs views.Materialized view in Teradata are called Join Index and can be found in the manuals using that term.
9226 Mar 2014 @ 01:02 PDTDatabaseReplyHelp with creating a Case within an In statment within a CaseYou need a separate Case on food type to display that differently. The two level case above only changes it inside that expression. Do you have permission to make a table of your own? Frequent cha...

Pages