|
Archives of the TeradataForumMessage Posted: Thu, 21 Dec 2007 @ 01:38:54 GMT
Another consideration is performance. Every time you access such a table function, it must: 1) Establish a connection to the remote system 2) Run a query 3) Transfer the results (over the LAN) to Teradata for inclusion in the query referencing the table function 4) Terminate the session. This happens each time you reference the table function. If you are running an insert select against the table function (i.e. transferring to Teradata) then using the newly populated table in subsequent queries, then that is OK. This means that the data is just transferred once. Accessing the table function in multiple queries is certainly workable for low volume data (whatever that means), but if you are talking large tables (e.g. customer, account, transaction, line_item) then you might find that sucking the data over the LAN each time you run a query (i.e. slow). If this is the case then you might find it easier to simply transfer the data using an external utility (e.g. export / fastload or tpump or ...) Glenn Mc
| |||||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | |||||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | |||||||||||||||||||||||||||||||||||||||||||||||||||