Unfortunately I can't help.
The difference in the structure of the table is easy to fix, but the compression algorithm remains unknown.
I already tried the various versions of the zenpxp algorithm and all the other algorithms without any luck.
P.S.: while doing my tests with one of the samples and using "comtype copy2" I guess I finally replicated the weird crash reported by some people here
viewtopic.php?p=65246#p65246 and
viewtopic.php?f=13&t=16455, and the good news is that the beta version isn't affected, so I was able to fix it in some way
