fix(bigquery): calculated column cannot orderby in BigQuery

superset/connectors/sqla/models.py

@@ -1300,6 +1300,13 @@ def get_sqla_query( # pylint: disable=too-many-arguments,too-many-locals,too-ma
  # if engine does not allow using SELECT alias in ORDER BY
  # revert to the underlying column
  col = col.element
+
+ if (
+ db_engine_spec.allows_alias_in_select
+ and db_engine_spec.allows_hidden_orderby_calculated_column
+ and col.name in [select_col.name for select_col in select_exprs]
+ ):
+ col = literal_column(col.name)
  direction = asc if ascending else desc
  qry = qry.order_by(direction(col))
 

superset/db_engine_specs/base.py

@@ -289,6 +289,12 @@ class BaseEngineSpec: # pylint: disable=too-many-public-methods
  # if TRUE, then it doesn't have to.
  allows_hidden_ordeby_agg = True
 
+ # Whether ORDER BY clause can use sql caculated expression
+ # if True, use alias of select column for `order by`
+ # the True is safe for most database
+ # But for backward compatibility, set to False by default
+ allows_hidden_orderby_calculated_column = False
+
  force_column_alias_quotes = False
  arraysize = 0
  max_column_name_length = 0

superset/db_engine_specs/bigquery.py

@@ -99,6 +99,8 @@ class BigQueryEngineSpec(BaseEngineSpec):
  # same cursor, so we need to run all statements at once
  run_multiple_statements_as_one = True
 
+ allows_hidden_orderby_calculated_column = True
+
  """
  https://www.python.org/dev/peps/pep-0249/#arraysize
  raw_connections bypass the pybigquery query execution context and deal with