Archives of the TeradataForum
Message Posted: Wed, 03 Aug 2011 @ 21:05:32 GMT
Subj: | | Re: Teradata Parallel transporter- Regarding schema |
|
From: | | Geoffrey Rommel |
| I am confused on what should be the schema to be used. | |
Yes, this can be confusing. You always need a schema unless you are *writing* to a database table. In this case, you can use a deferred schema
("schema *"), because TPT will pick up the schema from the database. In all other cases, you must provide one.
Load and Update operators: can use SCHEMA *.
Export operator: the schema describes the columns to be selected. This is needed because TPT doesn't know what columns you're going to select
or what their data types are; you might say "select *" or "select col1, col2, col3, ..." or "select case.....end ...".
Data Connector consumer: the schema describes the output file to be written. The data types are not necessarily the data types of the columns
as they exist in the database.
Data Connector producer: the schema describes the input file to be read.
|