On Thu, Aug 11, 2022, at 8:04 AM, Amit Kapila wrote:
>
> The reason to use text format is that it is error prone. There are restrictions
> while using the binary format. For example, if your schema has different data
> types for a certain column, the copy will fail.
>
Won't such restrictions hold true even during replication?
I expect that the COPY code matches the proto.c code. The point is that table
sync is decoupled from the logical replication. Hence, we should emphasize in
the documentation that the restrictions *also* apply to the initial table
synchronization.