New database or new family tree

I would like to get some input from other Gramps users about the question…

When should I store a new tree in a new database instead of a independent tree in the current database? Could there be a scenario that would advocate a new database instead of ‘just’ a new tree in the current instance?

It’s not an extreme large tree I’m thinking of (~ 200 ppl).

Thanks for any feedback

Gramps v 6.0.7 (flatpak) on Linux Mint/Ubuntu

Each family tree has its own database. So when you create a new family tree, you create a new sqlite database, and each database has its own subdirectory with a ‘sqlite.db’ file, that contains the data and indexes.
When you open a family tree in the Family Tree Manager, you create a connection to a specific database managed by SQLite.

True but I think @dbyy is asking whether or not he should have two ‘trees’ within the same db, i.e. having opened his existing tree, within that tree add a new person that has no connection to the existing tree and then develop a new tree around this isolated individual.

I realize that the connection between family tree and database is a little bit odd in Gramps (compared to other database’s (i.e Oracle or mySQL)).
(A) family tree(s) for i.e. in Oracle/mySQL would be in a set of tables with a common indexkey (for each tree). In Gramps is Familytree=database.

But you are right, basically I’m wondering, should I stuck several families (who are not related to each other) in one Tree (database) or should I create an independent database (in a possibly totally different directory path)? What are your thought about this?
Thanks

The only drawback to having multile family tree, each in their own database, is that you would need to create place records in each database. If there is no major overlap of places common to different trees this should pose no problem. But if many places are in multiple trees you would be duplicating your efforts.

An alternative is to create the multiple trees in the same database, or a copy of your main database. You could use the same place and source records for all trees. Once entered, you can export each tree for import into their unique database. Place and source records would export with each tree.

Good idea and a smart approach.