|
|
Archives of the TeradataForum
Message Posted: Mon, 03 Jul 2006 @ 11:54:32 GMT
Subj: | | Re: Basic questions about TPUMP |
|
From: | | McCall, Glenn David |
| Does TPump only work from a file, or can the data being taken from some other input ? | |
The basic tpump language allows you to specify a file. The file can be any type of file the host operating system supports (including a named
pipe fifo on Unix). You may also use access modules such as the MQ (to read data from a message queue), or the named pipes access module (which is
better than accessing a named pipe directly as per my previous statement). Finally, you can create an INMOD which can process records in any form
that you are capable of programming. For example, you can use an INMOD to receive data from an on-line service if you liked.
| If the data is taken from a file, how is data normally get cleared down from that file ? | |
I'm not sure what you mean by this. When tpump reads from a file, it will start at the first record and read sequentially through to the end.
It doesn't modify the content of the file. You can modify this with the "for" keyword. The for keyword allows you to control where tpump starts
reading and how many records are read.
| If I wanted to run a TPump job only once per day - basically working on a file that is produced once per day - how would you set up
the parameters for that sort of job ? | |
Basically you would schedule the tpump job via a scheduler. Obviously, you would need to ensure that whatever creates the data for tpump to
load has completed. If you need variable names for the file, then look at some of the other posts in the archive that detail how to do this with
host variable substitutions.
| Why is it necessary to have 'END LOADING' on a TPump script when it is a continuous feed ? Syntax, or is it more subtle than
that? | |
Think of the "END Loading" statement as a command that tells tpump that there are no more commands in its command file and that it can actually
begin processing the data.
| |