- is produced by a program that walks the "white tree." The white tree is the output of the o..." />
Archives of the TeradataForum
Message Posted: Mon, 17 Feb 2004 @ 01:33:10 GMT
What you're getting is two different views of the same basic plan. This is because the explains were produced by different programs. The "regular" explain - the one you get when you say explain <request> - is produced by a program that walks the "white tree." The white tree is the output of the optimizer phase of the parser. The "runtime" explain - the one you get from the monitor - is produced by a program that walks through steps the dispatcher has for the query. The final phase of the parser is called the Gen phase. What it does is translate the white tree into a series of step messages which get bundled up and sent to the dispatcher. The dispatcher then coordinates the execution of the plan by sending the messages to the amps.
The parser and the dispatcher used to be in different "partitions." Starting in V2R5 (I think), they're in the same partition. This will make it easier to iron out the differences between the two explains and I think we probably will eventually. I don't think they'll ever be exactly the same, though. When you get a runtime explain, the white tree for it is long gone, and I doubt that will ever change.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|