sadpandajoe commented on code in PR #43527:
URL: https://github.com/apache/superset/pull/43527#discussion_r3867651457
##########
superset/datasource/api.py:
##########
@@ -523,6 +556,255 @@ def compatible(self, datasource_type: str, datasource_id:
int) -> FlaskResponse:
return self.response(200, result=result)
+ def _resolve_for_query(
+ self, datasource_type: str, datasource_id: int, payload: dict[str, Any]
+ ) -> ResolvedExplorable:
+ """Resolve + authorize, translating DAO/security errors to HTTP."""
+ if DatasourceType(
+ datasource_type
+ ) == DatasourceType.SEMANTIC_VIEW and not
is_feature_enabled("SEMANTIC_LAYERS"):
+ raise _HttpError(404, "Semantic views are not enabled.")
+ try:
+ return resolve_explorable(
+ datasource_type,
+ datasource_id,
+ time_column=payload.get("time_column"),
+ has_time_range=bool(payload.get("time_range")),
+ )
+ except DatasourceTypeNotSupportedError as ex:
+ raise _HttpError(400, ex.message) from ex
+ except DatasourceNotFound as ex:
+ raise _HttpError(404, ex.message) from ex
+ except SupersetSecurityException as ex:
+ raise _HttpError(403, ex.message) from ex
+ except TabularQueryValidationError as ex:
+ raise _HttpError(400, str(ex)) from ex
+
+ @expose("/<datasource_type>/<int:datasource_id>/query", methods=("POST",))
+ @protect()
+ @safe
+ @statsd_metrics
+ @event_logger.log_this_with_context(
+ action=lambda self, *args, **kwargs:
f"{self.__class__.__name__}.query",
+ log_to_statsd=False,
+ )
+ def query(self, datasource_type: str, datasource_id: int) -> FlaskResponse:
+ """Query a datasource using metric and dimension names.
+ ---
+ post:
+ summary: Query a datasource by its semantic definitions
+ parameters:
+ - in: path
+ schema:
+ type: string
+ name: datasource_type
+ - in: path
+ schema:
+ type: integer
+ name: datasource_id
+ requestBody:
+ required: true
+ content:
+ application/json:
+ schema:
+ $ref: '#/components/schemas/DatasourceQuerySchema'
+ responses:
+ 200:
+ description: Query result
+ content:
+ application/json:
+ schema:
+ type: object
+ properties:
+ result:
+ type: array
+ items:
+ type: object
+ application/vnd.apache.arrow.stream:
+ schema:
+ type: string
+ format: binary
+ 400:
+ $ref: '#/components/responses/400'
+ 401:
+ $ref: '#/components/responses/401'
+ 403:
+ $ref: '#/components/responses/403'
+ 404:
+ $ref: '#/components/responses/404'
+ """
+ # No ValueError guard here: marshmallow raises ValidationError (not a
+ # ValueError subclass), and datasource_type comes from the path, not
the
+ # body. DatasourceType() is converted in _resolve_for_query below.
+ try:
+ payload = DatasourceQuerySchema().load(request.json or {})
+ except ValidationError as ex:
+ return self.response_400(message=ex.messages)
+
+ try:
+ resolved = self._resolve_for_query(datasource_type, datasource_id,
payload)
+ except _HttpError as ex:
+ return self.response(ex.status, message=ex.message)
+ except ValueError:
+ return self.response(
+ 400, message=f"Invalid datasource type: {datasource_type}"
+ )
+
+ if errors := validate_query_names(
+ resolved.valid_metrics,
+ resolved.valid_dimensions,
+ metrics=payload["metrics"],
+ dimensions=payload["dimensions"],
+ filters=payload["filters"],
+ order_names=[term["column"] for term in payload["order"]],
+ ):
+ return self.response_400(message="; ".join(errors))
+
+ return self._execute_and_respond(resolved, payload)
+
+ def _execute_and_respond(
+ self, resolved: ResolvedExplorable, payload: dict[str, Any]
+ ) -> FlaskResponse:
+ """Run the query and render it in the requested result format."""
+ grain_column = resolved.resolve_grain_column(
+ payload["time_column"], payload["dimensions"]
+ )
+ if payload["time_grain"] and not grain_column:
+ # Silently dropping the grain would return unbucketed rows that
look
+ # correct, so refuse instead.
+ return self.response_400(
+ message=(
+ "time_grain requires a temporal column. Set time_column, "
+ "include a datetime dimension, or provide time_range."
+ )
+ )
+
+ query_dict = build_query_dict(
+ time_column=resolved.time_column,
+ metrics=payload["metrics"],
+ dimensions=payload["dimensions"],
+ filters=payload["filters"],
+ time_range=payload["time_range"],
+ time_grain=payload["time_grain"],
+ grain_column=grain_column,
+ limit=payload["limit"],
+ offset=payload["offset"],
+ order=[(term["column"], term["descending"]) for term in
payload["order"]],
+ # No wire field for this; follow the leading term's direction.
+ order_desc=(
+ payload["order"][0]["descending"] if payload["order"] else True
+ ),
+ )
+ result_format = payload["result_format"]
+
+ try:
+ result = execute_tabular_query(
+ int(resolved.explorable.id),
+ str(resolved.explorable.type),
+ query_dict,
+ result_format=result_format,
+ use_cache=payload["use_cache"],
+ force=payload["force"],
+ cache_timeout=payload["cache_timeout"],
+ )
+ except SupersetSecurityException as ex:
+ return self.response(403, message=ex.message)
+ except (TabularQueryValidationError, QueryObjectValidationError) as ex:
+ return self.response_400(message=str(ex))
+ except CommandException as ex:
+ return self.response_400(message=ex.message or str(ex))
+
+ queries = result.get("queries") or []
+ if result_format == ChartDataResultFormat.ARROW:
+ if len(queries) != 1:
+ return self.response_400(
+ message="Arrow result format supports exactly one query."
+ )
+ return Response(
+ queries[0]["data"],
+ mimetype="application/vnd.apache.arrow.stream",
+ )
+ return self.response(200, result=queries)
Review Comment:
The JSON branch sends the query payload straight through the API JSON
response helper, but DataFrame records can still contain `pd.NaT`, timestamps,
or NaN values. Nullable temporal data can therefore raise during serialization,
while other values diverge from chart-data's normalized JSON representation.
Could this use the same JSON serialization/sanitization path as chart-data
before returning the result?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
