As you know, Isar works on mobile devices and desktops running on the VM as well as Web. Both platforms are very different and have different limitations.
- Only the first 1024 bytes of a string can be used for a prefix where-clause
- Objects can only be 16MB in size
Because Isar Web relies on IndexedDB, there are more limitations, but they are barely noticeable while using Isar.
- Synchronous methods are unsupported
.matches()filters are not yet implemented
- Schema changes are not as tightly checked as in the VM so be careful to comply with the rules
- All number types are stored as double (the only js number type) so
@Size32has no effect
- Indexes are represented differently so hash indexes don't use less space (they still work the same)
col.deleteAll()work correctly but the return value is not correct
col.clear()do not reset the auto-increment value
NaNis not supported as a value