There is a bug in pages processor.
If PLAIN-encoding page occurs after we have read dictionary page (and stored dictionary), the dictionary is unconditionally used, which leads to emty (undefined) values.
This PR fixes such a case.
Also, column with originalType = 'UTF8' now consistently converts to string, not only the dictionary-encoded values.
There is a bug in pages processor. If PLAIN-encoding page occurs after we have read dictionary page (and stored dictionary), the dictionary is unconditionally used, which leads to emty (undefined) values. This PR fixes such a case. Also, column with originalType = 'UTF8' now consistently converts to string, not only the dictionary-encoded values.