Archives of the TeradataForum

Message Posted: Wed, 29 Nov 2006 @ 23:53:14 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

<-- Anonymously Posted: Wednesday, November 29, 2006 18:52 -->

David,

If you are using name pipe from sqlplus to fastload, did you define the fastload to read data as delimited data set? If not, default is fastload format as result your system may be struggle try to convert the best it could.

If you are using Korn shell anyway, why not land the sqlplus output file in the server and compress the file then ftp or whatever other method you have to transfer to the TERADATA server. Uncompress the compressed file then load using delimited data in your fastload script. This ensure minimum dependency on the network - some.

Name pipe depends a lot on the network performance and processing power on both servers in either end of that network. I have noticed that AIX does not handle name pipe as efficiently as I would expect with Solaris. I am not sure what you have on both sides of the network so I can't really comment on this.

You could use the suggested method in second paragraph as your baseline and determine where the bottleneck is then address the issue accordingly.

Hope this helps.

	<Prev	Next>		<<First	<Prev	Next>	Last>>