I am trying to figure out how to take some tables and turn them into star schemas. My simplest example is a table that is keyed by a date/time field (containing both a date and a time) and then several quantity fields. These quantities are recorded every few minutes.
Kimball says on page 245 of "The Data Warehouse Lifecycle Toolkit" to split the date/time field up into two separate fields, one for date and one for time. I did this and was able to join my resulting fact table to my date dimension (day level).
Now my question is what do I do with the time field? My users are going to want to report these quantities by hour or maybe half hour. Do I also need a time dimension down the second? Kimball says to leave it in the fact table as an integer, but that won't help my users. Any advice?
Kimball says on page 245 of "The Data Warehouse Lifecycle Toolkit" to split the date/time field up into two separate fields, one for date and one for time. I did this and was able to join my resulting fact table to my date dimension (day level).
Now my question is what do I do with the time field? My users are going to want to report these quantities by hour or maybe half hour. Do I also need a time dimension down the second? Kimball says to leave it in the fact table as an integer, but that won't help my users. Any advice?