icexelloss commented on code in PR #35514:
URL: https://github.com/apache/arrow/pull/35514#discussion_r1213328400
##########
python/pyarrow/src/arrow/python/udf.cc:
##########
@@ -234,6 +351,56 @@ Status RegisterTabularFunction(PyObject* user_function,
UdfWrapperCallback wrapp
wrapper, options, registry);
}
+Status AddAggKernel(std::shared_ptr<compute::KernelSignature> sig,
compute::KernelInit init,
+ compute::ScalarAggregateFunction* func) {
+
+ compute::ScalarAggregateKernel kernel(std::move(sig), std::move(init),
AggregateUdfConsume, AggregateUdfMerge, AggregateUdfFinalize,
/*ordered=*/false);
+ RETURN_NOT_OK(func->AddKernel(std::move(kernel)));
+ return Status::OK();
+}
+
+Status RegisterAggregateFunction(PyObject* agg_function, UdfWrapperCallback
agg_wrapper,
+ const UdfOptions& options,
+ compute::FunctionRegistry*
registry) {
+ if (!PyCallable_Check(agg_function)) {
+ return Status::TypeError("Expected a callable Python object.");
+ }
+
+ if (registry == NULLPTR) {
+ registry = compute::GetFunctionRegistry();
+ }
+
+ static auto default_scalar_aggregate_options =
compute::ScalarAggregateOptions::Defaults();
+ auto aggregate_func = std::make_shared<compute::ScalarAggregateFunction>(
+ options.func_name, options.arity, options.func_doc,
&default_scalar_aggregate_options);
+
+ Py_INCREF(agg_function);
Review Comment:
My current probably incorrect understanding is that if we are passing a
PyOjbect to a OwnedRef or OwnedRefNoGIL then we need to call Py_INCREF because
the de constructor of OwnedRef and OwnedRefNoGIL will Py_XDECREF?
--
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]
