My current task is to create a report on the collection of a fairly large set of linear data (i.e. where the value 10 depends on some average factors from 9 8 7 6 5 4 3, etc.
The problem also is that this logic (now called its moving average) is not fixed and is likely to change. I need to not only generate the report in HTML, but also in Excel and, possibly, send it by e-mail for automatic use (so some of the variables will be relative, i.e. Generate a report for the last month, for example)
So my question is should I use a reporting mechanism or not? (Jasper, Pentaho, Burt)
My drivers are here - The reporting logic will change, but it will be changed by the developers. This will be complex logic, so it will be easiest to do in Java code. The report must be saved and sent by email in Excel format. The reporting challenge will be overwhelming (some things that may be useful to ETLs) - The requirements for future reporting are not covered here, the system will go over to rewriting if it should be general and expandable.
I could just create an object called report, save its type (enumeration) and a list of parameters (from the rest of the URL) and other metadata (who created what date, etc., where to send and relative parameters are saved) And then, when some processing was performed, save it in the database, just to get processed data for use by the controller. This greatly simplifies the technologies used (I use spring mvc, and this should be enough with a small cover of POI). But I wonder if it will be easier to use a reporting mechanism.
source
share