Dataflow GZIP TextIO ZipException: too many length or distance symbols

Looking at the code that produces the error message it seems to be a problem with zlib library (which is used by JDK) not supporting the format of gzip files that you have.

It looks to be the following bug in zlib: Codes for reserved symbols are rejected even if unused.

Unfortunately there's probably little we can do to help other than suggest producing these compressed file using another utility.

If you can produce a small example gzip file that we could use to reproduce the issue, we might be able to see if it is possible to work around somehow, but I wouldn't rely on this to succeed.