Mapped identifiers must be stored in an extended Babelfish catalog

[robverschoor]

Babelfish modifies certain identifiers in order to map T-SQL names to PG-compatible names. Examples include index names (which get a hash appended to make them unique in a schema) and identifiers longer than 63 characters (which get truncated at 31 and have a 32-char MD5 hash of the original full name attached). These identifiers should be recorded into the Babelfish extended catalog sys.babelfish_namespace_ext. The purpose of recording these identifiers is to enable user-facing communication, for example in error messages as well as in the contents of the T-SQL catalogs, to report the original T-SQL name instead of the internal Babelfish name.