24 | 06 Jun 2006 @ 01:21 PDT | Database | Reply | why this error was raised? help me! | A correlated subquery might be faster. Try this:update TTEMP.TB_MID_NEW_INNET_SUBSCRIBERset Active_Month = 200604where Active_Month is nulland substr(cast(START_DATE as format 'YYYYMM'), 1, 6) = '... |
23 | 05 Jun 2006 @ 02:47 PDT | UDA | Reply | Teradata Upgrade | Every new release contains many bug fixes, performance/availability/usability improvements, and new features. Consult the release documentation for details. |
22 | 05 Jun 2006 @ 01:13 PDT | Database | Reply | why this error was raised? help me! | Try this:update TTEMP.TB_MID_NEW_INNET_SUBSCRIBERfrom (select distinct MSISDN from PVIEW.VW_NET_GSM_NL_200604) aset Active_Month = 200604where Phone_No = a.MSISDN and Active_Month is nulland substr... |
21 | 05 Jun 2006 @ 01:00 PDT | UDA | Reply | Unloads from DB2 (nullable fields) - How to manage this with Fastload? | The NULLIF method doesn't work for numeric data types. To capture null values properly, you will need to convert your data to the kTeradata indicator format. |
20 | 01 Jun 2006 @ 01:32 PDT | Database | Reply | Difference between Teradata Utilities and ETL Process | This Wikipedia article may be helpful: http://en.widipedia.org/wiki/Etl |
19 | 31 May 2006 @ 03:58 PDT | Database | Reply | AMP on Teradata is UTILITY/DOWN | I assume that by "doing a REBOOT on this AMP" you mean that you used the BOOT command of Vprocmanager. The purpose of that command is prepare the AMP for a complete rebuild. Therefore, it erases ... |
18 | 25 May 2006 @ 11:29 PDT | Database | Reply | CLOB in Stored Procedure is not working | Evidently, it should have failed to compile due to the duplicate definition of "myclob". This appears to have been fixed recently, but I don't know which release has the fix. |
17 | 19 May 2006 @ 05:09 PDT | Database | Reply | Data Distribution | It all depends on how you define the primary index. You can place the rows all on one AMP, or you can distribute them randomly. If the table will never have more than a few rows, it would make se... |
16 | 04 May 2006 @ 10:46 PDT | Tools | Reply | How to display all columns with betq command | Use the ".set foldline on" command to allow BTEQ to fold a row across multiple lines. For a prettier format, also ".set sidetitles on". |
15 | 04 May 2006 @ 10:35 PDT | Database | Reply | How to find a lost password for a user | By design, passwords are stored in DBC.DBase using a one-way encryption method. If you forget a password, consider it gone forever. |
14 | 19 Apr 2006 @ 06:50 PDT | Database | Reply | Write/Read external file from within Stored Procedure? | There is no good way to do that with Teradata. Even if methods for accessing the host OS file system were available, it probably would not be useful on MPP configurations, since you don't have con... |
13 | 05 Apr 2006 @ 01:34 PDT | Tools | Reply | Numeric overflow | It may be worth mentioning that the upcoming release V2R6.2 will support up to 38 digits for the DECIMAL data type. |
12 | 30 Mar 2006 @ 11:51 PST | Database | Reply | "wait" operation in a Stored Procedure | I suppose a UDF could be written to perform this function. However, your application might be better served using a queue table. A queue table is a table that can be used as a FIFO event queue. ... |
11 | 27 Mar 2006 @ 02:48 PST | Database | Reply | number of CPUs per node | I think that method would work only if RSS logging is turned on. I suppose the most general method would be to look up the hardware specs for your server nodes. If those are not available for som... |
10 | 13 Mar 2006 @ 12:55 PST | Database | Reply | Table sizing in bytes | For an existing, populated table, you can query DBC.TableSize.Example:Select sum(currentperm) from dbc.tablesize where databasename = 'mydbname' and tablename = 'mytablename'; |
9 | 28 Feb 2006 @ 12:20 PST | Tools | Reply | Partitioned tables: multiload vs JDBC | Partitioning has no effect on locking granularity, so concurrency is the same. There is no partition level locking.Multiload always locks the entire table. Single row inserts and updates (assumi... |
8 | 28 Feb 2006 @ 12:09 PST | Database | Reply | Locking modifer for a MSR | The two requests are equivalent, so the execution plans are identical as shown by the explains. What difference were you expecting? |
7 | 24 Feb 2006 @ 02:28 PST | Database | Reply | Negative spool while monitoring Session | Are you using a very old version? Negative spool space problems were supposedly fixed a long time ago (V2R3.1). |
6 | 10 Feb 2006 @ 02:23 PST | Database | Reply | Tools to compare two teradata's DB and find the differences | Do you want to compare the contents of tables, or are you just looking for schema differences? |
5 | 08 Feb 2006 @ 03:35 PST | UDA | Reply | Deadlock and isolation levels | If your application queries a table and then updates it within the same transaction, it is vulnerable to a classic deadlock scenario. The query will acquire a read lock and then attempt to upgrade... |
4 | 02 Feb 2006 @ 11:29 PST | Database | Reply | Delete / truncate command | No, DROP TABLE is very fast, regardless of table size. It does not log individual rows. |
3 | 26 Jan 2006 @ 11:30 PST | Database | Reply | Delete / truncate command | In general, the unqualified delete statement uses a "superfast" method similar to what you described for the truncate command (and essentially the same as for drop table) whenever it can do so with... |
2 | 04 Jan 2006 @ 03:42 PST | Database | Reply | Cancel online transaction recovery after break big delete operation | You can use the CANCEL ROLLBACK ON TABLE command of the Recovery Manager utility. I believe this is available in 5.1 and later releases. If you use this command, the table is marked unusable, and... |
1 | 03 Jan 2006 @ 03:48 PST | Database | Reply | Gateway ports | All Teradata clients connect to TCP port 1025. That is the only port that you should need to forward through your firewall. |