Rum Database Structure

Hello there ! :slight_smile:

Here is a little topic to discuss with y’all rum lovers :slight_smile: After a brief talk with @Oliver, we felt that the rum database structure could be improved, since the rum classification can be very tricky sometimes.

It indeed requires to take into consideration some elements :

  • if you classify rums by distillery (which is according to me to most instinctive approach for the classification) then how about the lost distilleries from the past : for instance Cambridge is long gone but the style is still reproduced thanks to Long Pond, so can we consider it as a Long Pond due to the distillery, or as a Cambridge for the style ?

  • Sometimes the approach needs to go deeper, because a distillery can produce different rums for different brands. In Martinique for instance, Saint James produces J Bally rhums, or the Distillerie du Simon produces Clément, HSE and Braud & Quennesson rhums, which by the distillery approach results in a messy classification with rhums from all the brands inside the same category.

  • Sometimes the information can be misleading due to the bottler’s marketing : we all know for instance that Velier’s Versailles 1998 is not distilled or made at Versailles’ distillery since it’s long gone by that time.
    Why most Enmore made Versailles Still are categorized as « Enmore » but the Enmore 1994 from Nobilis is classified as « Enmore (Versailles) » unlike most of the other 1994 REV bottlings ? Of course, Guyana is a very tricky issue on that topic that could also be addressed.

Anyway, just wanted your thoughts on the database structure to make it easier and more convenient for us all !

Rumely yours :tumbler_glass:


Thanks for the good suggestions! Indeed, it seems relatively tricky to carefully categorize all releases.

I have some thoughts on this. If I’m missing the point, please correct me.

To your first point:
In this case, I think the classification “Long Pond” would definitely be correct, provided you search by distillery. “Cambridge” would clearly be incorrect, provided it is a Cambridge-style rum from the Long Pond distillery. Here, it would then rather require a “Style” category. In this example, I find it quite clear.

To your second point:
In the case of J. Bally, the distillery would in my opinion be Saint James and not J. Bally, because it is not a distillery. In this case, J. Bally would be classified as a bottler or publisher.

To your third point:
In the case of Guyana rums, I believe the same procedure should be followed as for point one. I would also categorize rums from closed distilleries under “distillery” with their original name. The actual bottler or publisher would again be listed separately.

I hope I have understood and categorized your points correctly.

However, I see another difficulty: for example, the Equiano Afro-Carribean Rum (RX4754) should be categorized under two distilleries, Foursquare and Grays. Currently it is classified under Grays, which in my opinion is insufficient for the search function or a clear categorization. Here it would make sense to specify several distilleries in order to find the rum both when searching for Foursquare and Grays.

Otherwise, I find the current database structure already very successful and helpful. I have not had any problems so far. Only multiple, identical rums I have already found, for example, when searching for Saint James Single Casks.

In any case, thanks for the diligent improvement of the database!


Wouldn’t it be a good solution to split style and distillery, ie “Versailles” (1998 for example) distilled at “Enmore”?


I got the point, people will come here looking for Skeldon and will find nothing because even the Velier’s Skeldons were distilled in Diamond.

I would suggest a tree structure. To my knowledge, no marque has ever migrated to another country. Therefore I would use country as primary category, distillery (so Long Pond, not Cambridge and then marque. There is plenty of room for supplementary information, like
"Note that the Cambridge rum style is reproduced by the Long Pond distillery since… ", Which can be added in the rum description.

Yeah, and of course, blends are different story. Imo it should be marked, whether we are dealing with single marque, single estate or multiple estates.

1 Like

Could be a solution. The only question is whether the benefit is in proportion to the effort required to implement something like this. Theoretically, the dataset would also have to be provided with the corresponding information retroactively, right? @Oliver How do you see this?

For me such a category was not really relevant so far. In most cases, one knows the style of the respective rum.


To be honest, I often go to rumx to verify things like this so it wouldn’t be bad if we have solid information :slight_smile:


So there seems to be a usecase :grinning:

I also like to have as much information as possible concentrated. I’m just wondering how much effort that would take and if it’s feasible. However, I don’t have a clue about app programming either.

If you really want complete info about the origin, I think you would need the categories “Distillery”, “Bottler” or “Publisher” and “Style”.

The first two each with the option to enter multiple details.


From the implementation side, we don’t have any (major) restrictions here. If we come up with a great customized model here, I can implement that in the app.

The bigger question is how we handle data maintenance. All existing Rums would have to be adapted to the new model. Some fields could certainly be adjusted automatically based on rules, but especially when new fields are added we would need to populate them community-driven and maintain the data.


This is exactly what I meant by the question about the cost-benefit ratio. It sounds like an extremely time-consuming measure to provide the complete database with the corresponding data.