First of all, to keep the most recent records in each group, it would be convenient to use a timestamp for each record. Another possibility would be to simply use the recno() to figure out the latest records; however, with a timestamp, it should be easier to create relevant queries; also, it would be scalable to a C/S environment.
An idea would be to use queries to select the data in the jobs table. Then, with the primary keys, you can fetch data from the related tables.
Another possibility is to delete all data that is NOT to be included in the "light" version. This might be simplified through RI.
>Hi everyone,
>
>We have a big database with ~ 20 related tables inside. There are no triggers or rules, as I know. One of the tables is called Jobs and has JobID (I), JobType, LastRunDate, etc. fields. This table has several dependent tables, such as JobSteps, for example. I want to have a light version of that database with two most recent jobs per each job type. How can I automate the process of creating such "light" version of a database? And I am also curious, if a generic tool, which allows to easily do such task, already exists.
>
>Thanks a lot in advance.
Difference in opinions hath cost many millions of lives: for instance, whether flesh be bread, or bread be flesh; whether whistling be a vice or a virtue; whether it be better to kiss a post, or throw it into the fire... (from Gulliver's Travels)
