When writing the endpoints in java, for finding items by their keys, should I use the Id or the webSafeString of the key? In what situations does this matter?
It's up to you.
Do the entities have parents? Then you probably want to use the
urlsafe representation as a single string will contain the full path to the entity. If you used an ID instead - you would somehow need to manually include the IDs of all parents up to the root.
No parents & IDs are numeric / alphanumeric? Then just use the IDs as they look cleaner (again, this is not a rule and is completely up to you).
No parents but IDs have special characters in them? Use the
urlsafe representation as you might have issues with not being able to use some special characters without encoding them in HTTP.
Note #1: the
urlsafe representation have the entity names encoded that can be easily decoded, this is unlikely a privacy issue but you still should be aware of it. The actual data (IDs) are also simply encoded and can be easily decoded, so be careful when you use personal information such as emails as IDs, they are not safe with
Note #2: if you decide to change the structure of your data in the future (parents <-> children), you might get stuck with some
urlsafe data you issued to your users who are not aware of the changes you might have done.