Snowflake join table(flatten(.)) parsed as anonymous table

[mattijsdp]

Fully reproducible code snippet
Please include a fully reproducible code snippet or the input sql, dialect, and expected output.

In Snowflake SQL it is possible to flatten objects in (at least) the following two ways: using a lateral flatten and using a join table flatten. If we then compute the lineage of a column taken from the flattened source, the catalog and db of the source is accounted for in the former but not the lateral. I think this is because the lateral is parsed as an Anonymous table.

sql_lateral = """
select f.value:external_id::string as external_id
from database_name.schema_name.table_name raw,
lateral flatten(events) f
"""
sql_join = """
select f.value:external_id::string as external_id
from database_name.schema_name.table_name raw
join table(flatten(events)) f
"""

lat_expression = parse_one(sql_lateral, dialect="snowflake")
join_expression = parse_one(sql_join, dialect="snowflake")

lat_root_node = lineage("external_id", lat_expression, dialect="snowflake")
join_root_node = lineage("external_id", join_expression, dialect="snowflake")

def print_source(root_node):
    for node in root_node.walk():
        source = node.source
        if isinstance(source, exp.Table):
            print(f"{source.catalog}.{source.db}.{source.this}")

print_source(lat_root_node) # DATABASE_NAME.SCHEMA_NAME.TABLE_NAME
print_source(join_root_node) # ..TABLE(EXPLODE(EVENTS))

Official Documentation
Please include links to official SQL documentation related to your issue.