Recognizes the Spark SQL pattern
explode(from_json(<column>, '<literal ARRAY<STRUCT<...>> schema>'))
(also
explode_outer) and produces a vendor-neutral
StructuredDataflowDescriptor. Unsupported variants
(dynamic schema column,
schema_of_json(...), MAP element fields,
posexplode,
inline, etc.) return
null so the
existing generic function lineage runs unchanged.
The same syntactic pattern is supported by Hive (which shares Spark's
from_json + explode UDFs and ARRAY/STRUCT type-string
grammar), so the adapter also matches EDbVendor.dbvhive.