There is an unsaved comment in progress. You will lose your changes if you continue. Are you sure you want to reopen the work item?
Suggestion: AS database is easily configured to contain realistic "current date" data
In certain types of demos or tests it's helpful if the AS database 1) has a member that really aligns to current server day/time and 2) only includes data processed through that day, no data beyond that day.
We understand that the design of AdventureWorks AS database is such that seasonality is realistic/explainable, transactions are realistic/explainable, and that certain trends are known in the dataset. But, it also is unrealistic in that it's completely static
& we have all said "pretend today is Jan 4 2003, & please ignore that this chart shows data for tomorrow, and for next month, and for Christmas... those haven't happened yet."
Would like to see:
1) That the start date for the entire set sample data set can be set by default (say, 2001) convention (current year) or configuration (pick a year.) Modulating only the year is adequate, because that preserves the seasonaility that's built into the sample
2) Then, that the entire data set is packaged in "future" and "past" tables, and a provided SSIS package can move the data from "past" into "future" to catch up to a particular date, or on a daily basis (or some other
configured timespan), & trigger cube incremental process. For example, if AW ships with 4 years of sample data, for example, you might configure it so 3 years are "past" and 1 year is "future", then run the SSIS package to catch "past"
up to current date, then schedule the SSIS package to run daily to provide a dynamic edge to the data set for <1 yr. Or, you might use 2 years of "past" and 2 years of "future" to get<2yrs of new data.