|
Archives of the TeradataForumMessage Posted: Tue, 01 Dec 2009 @ 19:34:11 GMT
Can anyone help me with some Sql, I wonder. Here's what I got so far .... 28/12/1997 1 1998 1 29/12/1997 2 1998 1 30/12/1997 3 1998 1 31/12/1997 4 1998 1 01/01/1998 5 1998 1 02/01/1998 6 1998 1 03/01/1998 7 1998 1 04/01/1998 8 1998 2 05/01/1998 9 1998 2 06/01/1998 10 1998 2 07/01/1998 11 1998 2 08/01/1998 12 1998 2 09/01/1998 13 1998 2 10/01/1998 14 1998 2 11/01/1998 15 1998 3 12/01/1998 16 1998 3 13/01/1998 17 1998 3 14/01/1998 18 1998 3 15/01/1998 19 1998 3 16/01/1998 20 1998 3 17/01/1998 21 1998 3 Column 1 : date Column 2 : generated nicely and easily Column 3 : Year Column 4: week within the Year Column 5 : (not shown) ... see text below The second column is simply generated (using ROW_NUMBER() OVER (ORDER BY DTE_DT ASC) AS DAY_NR), which will continue to increment for all days in the calendar, and continue to increment after 1998 - as required. Column four (which is week number), however, I want to use to provide another column (column 5) incrementing forever, but not resetting at the start of the new year. So, basically, I want to have a count which only changes by 1, on change of that fourth column, ordering by Year. Within 1998 it would actually look the same as column four but when we go to 1999, that column would continue to increment, but again only when the week number changes. I can do this in procedural code, but I suspect it can't be done in SQL. Don't knock yourself out over this one ... I've already done that ! I guess you either know how to do it or you don't. Thanks and Regards David Clough
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | ||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | ||||||||||||||||||||||||||||||||||||||||||||||||