Hi all,
I think I am at a point where I need to consider looking into Glide Big Tables for our company apps, at least for some tables. My main concern for doing so is the 2 Line Item tables that we use for Sales Order Items and Purchase Order Items. I am currently using relations and roll-ups on these tables as a base for my Reports dashboard to show me sales statistics. When it comes to reporting on just sales figures per month, year etc. I think I can work around any Big Table limitations by creating a separate table holding monthly reports for 1 year and then writing to a “Year” report table once a year.
But my struggle is how to be able to look at statistics for individual products. For example - “How many X did we sell between January and March 2025?”. “How much of our total monthly sale for month X in year Y was generated from product Z?”
In cases like this we sometimes go back and look for statistics 2-3 year to track sales for a specific item. I could do monthly count, yearly count columns on the actual product table, but I feel like there must be a better way? Some kind of JSON column in the products table storing data for that product? Or something else?
Any help and guidance is much appreciated 
It depends on the types of columns your rollups are referencing. If you’re referencing only non-computed columns, then there are essentially no limitations.
If there are computed columns in play, then it gets a little more complicated.
Some are not supported at all, some are some are supported but limited to 100 rows, and some are supported with no limitations.
It’s important to keep in mind that when you have multiple layers of computed columns, then the lowest common denominator determines the level of support.
Interesting! I might have misunderstood the limitations then. So let’s say each row in the line items table has a Product ID inserted in a non computed column. If I then in my Products table create a relation to the Product ID column in the Big Table, then that relation or query would return all the related rows from the Big Table without limitations?
Yes, it should.
Very easy to test, by the way.
- Create a Big Table
- Export the data from your existing regular table
- Import to the Big Table
- Reconfigure your relation/rollup to reference the Big Table.
1 Like
Brilliant, thank you very much for this. Since I have no experience when it comes to Big Tables at all, is there anything I need to keep in mind specifically when structuring my tables and data compared to building with normal Glide Tables? I understand that it is a difficult question to answer since each app is different. I was thinking more in general terms of any common workflows, practices that I might need to approach differently?
About the only thing I would say is to favour non-computed columns over relations+lookups as much as possible, and especially so if you intend using those looked up values as filters in Collections.
For example, you might normally store a UserID and then use a relation + lookup to get the User Name. If you wanted to then use the Name as a filter in a Collection, it won’t work. So in cases like that I tend to store both the UserID and User Name in non-computed columns. It goes against standard data normalisation practices, but sometimes is necessary.
3 Likes
Thanks, that is very helpful. I guess my best option might be to start building and testing to understand specifically what workarounds I might need to implement in each case. Really appreciate your helpful input on this.