Hi,


I have a large query where there are two window sets, one over year and another over quarter. I think that since teh parttion by and order by clause is same for 1 set of column and another set of columns, this query can be re-written to avoid the partioning on rows being done again and again. Below is the query:

I have a requirement to fit in 441GB of history data into a table which makes my table bulky and difficult to query. Just to explain it is a fact table and stores amount fields etc. I have a partition on the business date field.

Hi all,
We are experiencing a relative slow ODBC throughput from TD to Microstrategy and I am wondering if this is "as good as it gets" or if the throughput is below standard.

I have a transactional dataset which has a date field, my query is whether it is better to bring in the date field into the PI for the table i am using or leave it out. The PI is currently made up of a product unique identifier which is duplicated for each date related to the product.

Given: a large table; 14/50 fields responsible by one group; the rest by a second group; the need to for 5/14 fields to be used by analysts familar with the large table (ie. large numbers of scripts written against the current large table; business hour performance on this large table is a primary concern. Updates are typically schedule after hours.

Hi all,

I have a particular question regarding how the data are stored inside ResUsageSAWT.
When the column TheTime is for example '04:10:00.00' all the data regarding AWTs refers to the interval from 04:00 to 04:10 or refers to the interval 04:10 to 04:20?
Thanks in advance for your help!

Hi, I am new to teradata learning, I have download vm Express edition for testing purpose, but this is a very limited in performance, i want to test a machine have a little bit big data, ALL i have to face that, when ever i run a query especially with aggregation, i takes lots of mins to perform the resul

There is  "poetic query" that takes forever to run
Kind of looks like this
sel
col1.tb1,
col2,tb1,
(case <condition involving tb1 and tb2>) as "Dcol1",
Sum ( col1.tb3),
sum (col2.tb3 ),
sum (col3.tb3)
etc
from
tb1 left outer join tb2 <condition> LOJ tb3 <conditions>

Hi,
We have a set table with NUPI and is partitioned with date. My question is when we try inserting a new record based on row hash it will perform duplicate checking or it will go to particular partition and perform duplicate check?
Ex: in one AMP
Primary Index(ex: ID1): 10
Partition 1 with date 2015-01-01: 100 rows

Hi Guys, thanks for reading the post. I did do a search for "view on view" and there was no result.

Hi All,

I am trying to debug a query which is taking a lot of time.
Following objects are involved in the query :
 

Hi all. We use PDCR to archive our DBQL tables - the primary index on the tables is (logdate,procid,queryid) and they are partitioned by logdate.
 

Typical 2 years Sales Data Mart contains detail and aggregates in order to support time traverse metrics like 4 Week Shipments, 13 Week Shipments and so on... 

What is the difference between simple system view like dbc.tables, X views and V views and which one is better - performace wise?
Does every system view like dbc.tables will always have a corresponding X and V view? I failed to find any X or V view for for dbc.ErrorMsgs?
 
Thanks

Hi ,
Can anybody please tell me which one is better performance wise ... Dynamic or Static Query ?
I have a procedure full of dynamic queries which all can be written as static queries.
But, not sure if it is worth the effort.

For a project, we have several groups that will produce numbers using models. Input and output data will be in the DB. Each group is independent and works in their own different way; however, the outputs they produce will need to be combined, aggregated and queried. 

Hi Teradatares :-)
I am new to Teradata database. Currently I am thiking of creating "big" table that should grow 200 mln rows each week. After a year data it will be 10 000 000 000 rows
After reading some best practice, documentation and forum, that's what I kind of understand.
Create table test

Hi,
What happens to the performance when you load same amount of data into same table on two servers one with 140 vprocs and another with 70 vprocs ? Does both the cases take same time to laod or the one with highest vprocs take less time than the one with less vprocs ??
 
Thanks,
Kapil

Hi,
 
Is there any way to measure performance impact on existing queries with enabling LockLogger?
 
I'm planning to enable LockLogger in continuous mode to debug deadlock queries.
But I'm not sure the side-effect on perfomance of existing queries. Will is use CPU a alot? Many Disk I/O?
 

Hi Experts,
There is a tpump which is taking 10 hours to process 7000 records, whcih earlier was taking 1 hour..  not sure why its taking 
so long now. the script cotains the normal insert select statement. 
just want to know that is there any optimal parametrs setting for the various TPUMP parameters..
like 
PACK

Hi All,
    Is there anyway to improve the performance(Reducing the CPU consumed is the main aim for now) of queries involving like in the where clause.
 
 
WHERE FIRST_NM LIKE :V_FIRST_NM||'%'
AND LAST_NM LIKE :V_LAST_NM||'%';

I have few doubts regarding statistics collection.
 
1. Does the order of columns used in collect stats command important? 
2. Consider the following scenario 
 
Table A joins with Table B based on Col1 and Col2 and Table A joins with Table C based on Col1, Col2 and Col3.

Hi All,
I have two tables lets say Tab1 with 1million records and Tab2 with 100million records.
Option 1) sel columns from tab1 inner join tab2
Option 2) sel columns from tab2 inner join tab1
Is option one better than option two? does it make a difference?

Hi All
I want to understand if there will be a difference in performace or on any front between
Any inputs will be of a great help. I am dealing with huge fact tables and many volatile tables to be created in a stored procedure.
 

Does Sql have better performance if columns have COMPRESS? If so how does COMPRESS help the performance? Is it because of lesser Spool space?

Hi everybody,
Could anybody tell me how TD analytical functions are implemented under the hub?

I've got Fastload opening a named pipe, and I'm redirecting the output from a SQL*Plus script to the pipe in Windows. I assumed this would be far faster than allowing SQL*Plus to spool the result to a file locally, and fastload to import the flat file. Turns out it's not.

When I call

ResultSet rs = dbmd.getColumns(null, "CUSTDATA", "TRANS", null);

it takes 16 seconds to get an answer back on a fairly unloaded system.  This table has 27 columns.

This seems to be excesvily slow.  Is there anything that can be done to speed this up?

Hi,

could you please let me know whether there will be any difference in performance between the following two queries. As i don't have any internal knowledge about sql, i have this doubt.

table A: column alpha has 100 million records
table B: column beta has 45 million records 

Hi ALL, I am  new  to Teradata so sorry if my questions seem trivial and obvious for you.

http://www.taubkin.de/workshop.html

We are looking to do some application performance testing and need to simulate a "real world" environment in which the warehouse is under strain.  Are there any tools/scripts out there that folks use to simulate the load typically placed on an EDW where is it generally pegged at 100% all day?

Is it a good practice to define a Primary Index on VARCHAR column ? Although I dont see any strange (data type conversions) in explain plans between Join operations on VARCHAR (Primary Index) to CHAR column (Primary Index) since they are from same domain, I just want to understand if this is as efficient as Joins between CHAR to CHAR columns.