I have a reporting system with many entities. I will limit this to two, players and teams, in this question, but there are actually many more. Both players and teams have many plays. The reports are plays, such as "For player X, what percent of plays were for positive yards?" Plays can have multiple attributes, such as weather, quarter, field position, indoor/outdoor stadium, etc...

The overarching goal is asking what percent of plays were for positive yards based on a lot of criteria. What I currently do, without any form of database, is create a vast set of text files. Each entity has a text file. Each player has one text file containing the unique ID of every play the player participated in. Every team has a similar file. Then, there is a single file with the unique ID of every play that went for positive yards. Further, every attribute has a text file containing the unique ID of every play that meets that attribute. With these files, I can very quickly answer something like "For player X, when playing Team Y, under conditions A, B, C, what percent of plays went for positive yards?" I simply gather up the IDs that are shared between all the text files and count them.

This should be a data cube. But, in all of the data cube designs that I see, nothing looks anything like this format. I can't find anything that lets me add/remove attributes of the plays. I can't find anything that would let me limit the query to plays with Player X, Player Y and Team Z. So, I am asking for someone to point me in the proper direction so I can see examples that are somewhere similar to this. Once I see a little bit, I am sure that I can complete the rest of the design.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.