fix(eap): support integer dynamic columns within VCC

pyproject.toml

@@ -31,7 +31,7 @@ dependencies = [
     "sentry-arroyo>=2.38.7",
     "sentry-conventions>=0.3.0",
     "sentry-kafka-schemas>=2.1.24",
-    "sentry-protos>=0.7.0",
+    "sentry-protos>=0.8.9",
     "sentry-redis-tools>=0.5.1",
     "sentry-relay>=0.9.25",
     "sentry-sdk>=2.35.0",

snuba/web/rpc/v1/resolvers/R_eap_items/resolver_trace_item_table.py

@@ -145,16 +145,28 @@ def transform_expressions(expression: Expression) -> Expression:
             return expression
         context = mapped_column_to_context.get(str(expression.key.value))
         if context:
+            source_type = (
+                context.from_column_type
+                or NORMALIZED_COLUMNS_EAP_ITEMS.get(
+                    context.from_column_name, [AttributeKey.TYPE_STRING]
+                )[0]
+            )
+            assert source_type in (
+                AttributeKey.Type.TYPE_STRING,
+                AttributeKey.Type.TYPE_INT,
+            ), "VCC can only map string or int attributes"
             attribute_expression = attribute_key_to_expression(
                 AttributeKey(
                     name=context.from_column_name,
-                    type=NORMALIZED_COLUMNS_EAP_ITEMS.get(
-                        context.from_column_name, [AttributeKey.TYPE_STRING]
-                    )[0],
+                    type=source_type,
                 )
             )
             return f.transform(
-                f.CAST(f.ifNull(attribute_expression, literal("")), "String"),
+                if_cond(
+                    f.isNull(attribute_expression),
+                    literal(""),
+                    f.toString(attribute_expression),
+                ),
                 literals_array(None, [literal(k) for k in context.value_map.keys()]),
                 literals_array(None, [literal(v) for v in context.value_map.values()]),
                 literal(context.default_value if context.default_value != "" else "unknown"),
@@ -608,7 +620,10 @@ def _get_page_token(
                 # the routing strategy will properly truncate the time window of the next request
                 return FlexibleTimeWindowPageWithFilters.create(
                     request,
-                    TimeWindow(original_time_window.start_timestamp, time_window.start_timestamp),
+                    TimeWindow(
+                        original_time_window.start_timestamp,
+                        time_window.start_timestamp,
+                    ),
                     response,
                 ).page_token
     else:
@@ -677,10 +692,15 @@ def resolve(
         except Exception as e:
             sentry_sdk.capture_message(f"Error merging clickhouse settings: {e}")
         original_time_window = TimeWindow(
-            start_timestamp=in_msg.meta.start_timestamp, end_timestamp=in_msg.meta.end_timestamp
+            start_timestamp=in_msg.meta.start_timestamp,
+            end_timestamp=in_msg.meta.end_timestamp,
         )
         snuba_request = _build_snuba_request(
-            in_msg, query_settings, routing_decision.time_window, routing_decision.tier, self._timer
+            in_msg,
+            query_settings,
+            routing_decision.time_window,
+            routing_decision.tier,
+            self._timer,
         )
         res = run_query(
             dataset=PluggableDataset(name="eap", all_entities=[]),

tests/web/rpc/v1/test_endpoint_trace_item_table/test_endpoint_trace_item_table.py

@@ -1,6 +1,6 @@
 import random
 import re
-from datetime import datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from math import isclose
 from typing import Any
 from unittest.mock import MagicMock, call, patch
@@ -3471,6 +3471,105 @@ def test_multiply_attribute_aggregation(self) -> None:
         assert len(res.results) == 1
         assert isclose(res.results[0].val_double, expected_avg)
 
+    def test_occurrence_virtual_column_mapping(self) -> None:
+        """
+        Reproduces the request shape sent by api.organization-events for the issues
+        view: OCCURRENCE items with virtual column contexts that remap group_id ->
+        issue and sentry.project_id -> project / project.name.
+        """
+        org_id = 1
+        project_id = 1
+        item_ts = datetime.fromtimestamp(1773929000, tz=timezone.utc)
+
+        items_storage = get_storage(StorageKey("eap_items"))
+        write_raw_unprocessed_events(
+            items_storage,  # type: ignore
+            [
+                gen_item_message(
+                    start_timestamp=item_ts,
+                    type=TraceItemType.TRACE_ITEM_TYPE_OCCURRENCE,
+                    attributes={"group_id": AnyValue(int_value=1)},
+                    project_id=project_id,
+                    organization_id=org_id,
+                    remove_default_attributes=True,
+                )
+            ],
+        )
+
+        message = TraceItemTableRequest(
+            meta=RequestMeta(
+                organization_id=org_id,
+                referrer="api.organization-events",
+                project_ids=[project_id],
+                start_timestamp=Timestamp(seconds=1773925780),
+                end_timestamp=Timestamp(seconds=1773933040),
+                trace_item_type=TraceItemType.TRACE_ITEM_TYPE_OCCURRENCE,
+                downsampled_storage_config=DownsampledStorageConfig(
+                    mode=DownsampledStorageConfig.MODE_NORMAL,
+                ),
+            ),
+            columns=[
+                Column(
+                    key=AttributeKey(type=AttributeKey.TYPE_STRING, name="issue"),
+                    label="issue",
+                ),
+                Column(
+                    key=AttributeKey(type=AttributeKey.TYPE_INT, name="group_id"),
+                    label="group_id",
+                ),
+                Column(
+                    key=AttributeKey(type=AttributeKey.TYPE_STRING, name="project"),
+                    label="project",
+                ),
+                Column(
+                    key=AttributeKey(type=AttributeKey.TYPE_STRING, name="sentry.item_id"),
+                    label="id",
+                ),
+                Column(
+                    key=AttributeKey(type=AttributeKey.TYPE_STRING, name="project.name"),
+                    label="project.name",
+                ),
+            ],
+            limit=101,
+            page_token=PageToken(offset=0),
+            virtual_column_contexts=[
+                VirtualColumnContext(
+                    from_column_name="group_id",
+                    to_column_name="issue",
+                    value_map={"1": "BAR-1"},
+                    from_column_type=AttributeKey.TYPE_INT,
+                ),
+                VirtualColumnContext(
+                    from_column_name="sentry.project_id",
+                    to_column_name="project",
+                    value_map={str(project_id): "bar"},
+                ),
+                VirtualColumnContext(
+                    from_column_name="sentry.project_id",
+                    to_column_name="project.name",
+                    value_map={str(project_id): "bar"},
+                ),
+            ],
+        )
+
+        response = EndpointTraceItemTable().execute(message)
+
+        assert [c.attribute_name for c in response.column_values] == [
+            "issue",
+            "group_id",
+            "project",
+            "id",
+            "project.name",
+        ]
+
+        col = {c.attribute_name: c for c in response.column_values}
+        assert col["issue"].results == [AttributeValue(val_str="BAR-1")]
+        assert col["group_id"].results == [AttributeValue(val_int=1)]
+        assert col["project"].results == [AttributeValue(val_str="bar")]
+        assert col["project.name"].results == [AttributeValue(val_str="bar")]
+        assert len(col["id"].results) == 1
+        assert col["id"].results[0].val_str != ""
+
 
 def _str_array(*values: str) -> AnyValue:
     return AnyValue(array_value=ArrayValue(values=[AnyValue(string_value=v) for v in values]))