Archives of the TeradataForum

Message Posted: Mon, 11 Jul 2011 @ 14:53:06 GMT


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

Subj:		Re: Fast load in application phase

From:		David Wellman

Hi Sukul,

You're understanding is correct. Remember that with both FL and ML, the client system sends data in blocks. Each block will (typically) contain multiple rows. The block is sent to an AMP, which then breaks the block into individual rows which are then sent to the correct AMP. This process happens in 'phase 1' for FL and 'acquisition phase' for ML.

1) Correct, FL does NOT use any spool space.

2) Correct, ML sorts data, just like FL.

Cheers,

Dave

Ward Analytics Ltd - Information in motion (www.ward-analytics.com)