Obviously these may cause difficulties downstream if large amounts of state build up in a single row, but we should still make them available and deal with the problems as we find them.
Some ideas:
- Small lists using the atomic op append_if_fits
- Larger lists using versionstamps (each table row ends up being like a topic)
- Sets
- Sorted collections
Obviously these may cause difficulties downstream if large amounts of state build up in a single row, but we should still make them available and deal with the problems as we find them.
Some ideas: