Archives of the TeradataForum
Message Posted: Sat, 17 Jun 2000 @ 15:37:26 GMT
I've seen very little optimizer regression across several customers; however, this trick does seem like a good candidate for regression since it is pretty obtuse. The existence of a column seems most suitably answered by the DBC dictionary table DBC.columns rather than this approach. All columns are available in this view, but possibly not all information about the column like data type, etc...; So it really depends on why the tool is doing this request. If the request is to provide additional types of information like data type, the tool should either use the standard ODBC / CLI call or the help column command by table or column which should provide all of the information that you require. Also, is this a in-house built tool or standard publicly available product?
As usual the SQL, Table / View DDL, Statistics / Help Index for all Indexes,
as well as Row counts of the tables, and the explains for the SQL involved would help us understand where this problem falls.
This data collection task will soon be easier with the new functions in V2R4.
Finally, I still hesitate to push for any of the choices that you request. One exception is that I believe there is a system setting on how many look aheads and how much energy the optimizer uses to find the best plan.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|