nastra commented on code in PR #15448:
URL: https://github.com/apache/iceberg/pull/15448#discussion_r2904986928
##########
core/src/test/java/org/apache/iceberg/rest/TestRESTScanPlanning.java:
##########
@@ -953,4 +957,110 @@ public void serverSupportsPlanningButNotCancellation()
throws IOException {
// Verify no exception was thrown - cancelPlan returns false when endpoint
not supported
assertThat(cancelled).isFalse();
}
+
+ @ParameterizedTest
+ @EnumSource(PlanningMode.class)
+ void fileIOForRemotePlanningIsPropagated(
+ Function<TestPlanningBehavior.Builder, TestPlanningBehavior.Builder>
planMode) {
+ RESTCatalogAdapter adapter =
+ Mockito.spy(
+ new RESTCatalogAdapter(backendCatalog) {
+ @Override
+ public <T extends RESTResponse> T execute(
+ HTTPRequest request,
+ Class<T> responseType,
+ Consumer<ErrorResponse> errorHandler,
+ Consumer<Map<String, String>> responseHeaders,
+ ParserContext parserContext) {
+ T response =
+ super.execute(
+ request, responseType, errorHandler, responseHeaders,
parserContext);
+ return maybeAddStorageCredential(response);
+ }
+ });
+
+
adapter.setPlanningBehavior(planMode.apply(TestPlanningBehavior.builder()).build());
+
+ RESTCatalog catalog =
+ new RESTCatalog(SessionCatalog.SessionContext.createEmpty(), (config)
-> adapter);
+ catalog.initialize(
+ "test",
+ ImmutableMap.of(
+ CatalogProperties.FILE_IO_IMPL,
+ "org.apache.iceberg.inmemory.InMemoryFileIO",
+ RESTCatalogProperties.REST_SCAN_PLANNING_ENABLED,
+ "true"));
+
+ Table table = restTableFor(catalog, "file_io_propagation");
+ setParserContext(table);
+
+
assertThat(table.io().properties()).doesNotContainKey(RESTCatalogProperties.REST_SCAN_PLAN_ID);
+
+ TableScan tableScan = table.newScan();
+ assertThat(tableScan.io().properties())
+ .isSameAs(table.io().properties())
+ .doesNotContainKey(RESTCatalogProperties.REST_SCAN_PLAN_ID);
+ // make sure remote scan planning is called and FileIO gets the planId
+ assertThat(tableScan.planFiles()).hasSize(1);
+
assertThat(table.io().properties()).doesNotContainKey(RESTCatalogProperties.REST_SCAN_PLAN_ID);
+
assertThat(tableScan.io().properties()).containsKey(RESTCatalogProperties.REST_SCAN_PLAN_ID);
Review Comment:
> Why can't we return null before the FileIO is available?
We expect `FileIO` to be non-null across the codebase, so returning `null`
seems counter-intuitive here. But I think I get your point: Currently with
remote planning the Scan's FileIO is actually only usable after calling
`planFiles`. Before that it's (depending on permissions) unusable and would
lead to permission issues
--
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]
