The PI System Connector 3 uses the "compressing" attribute of tags on the source PI Data Archive to determine if compression should occur when it transfers data to the destination PI Data Archive. However, the PI Buffer Subsystem of any PI Interfaces that feed data to the source PI Data Archive also uses the "compressing" attribute of these same tags to determine if compression should occur. It is possible for compression to occur for the Pi Buffer Subsystem but not for the PI System Connector 3 if the tag is part of the snapshot query, but it is impossible for compression to occur for the PI System Connector 3 but not the PI Buffer Subsystem.
I might be getting some of the details wrong since the PI System Connector 3 is still new to me (or everyone at this point), but my point is that the "compressing" attribute is conventionally used by programs that input data to a PI tag and not programs that read data from a PI tag, and the exception is the PI System Connector 3, and so the "compressing" attribute is used by multiple programs and cannot be set differently for each program.
I feel like the PI System Connector 3's use of the "compressing" attribute should be replaced with something like a configuration file and a configurable default setting for compression if a PI tag is not specified in the configuration file.
If I am completely missing the mark and this design choice actually makes sense, please let me know. I will admit that this is all a bit confusing to me.