diff --git a/application_sdk/observability/logger_adaptor.py b/application_sdk/observability/logger_adaptor.py
index 3f719a785..2dade2993 100644
--- a/application_sdk/observability/logger_adaptor.py
+++ b/application_sdk/observability/logger_adaptor.py
@@ -7,9 +7,9 @@
 from typing import Any, Dict, Optional, Tuple
 
 from loguru import logger
-from opentelemetry._logs import SeverityNumber
+from opentelemetry._logs import LogRecord, SeverityNumber
 from opentelemetry.exporter.otlp.proto.grpc._log_exporter import OTLPLogExporter
-from opentelemetry.sdk._logs import LoggerProvider, LogRecord
+from opentelemetry.sdk._logs import LoggerProvider
 from opentelemetry.sdk._logs._internal.export import BatchLogRecordProcessor
 from opentelemetry.sdk.resources import Resource
 from opentelemetry.trace.span import TraceFlags
@@ -495,7 +495,6 @@ def _create_log_record(self, record: dict) -> LogRecord:
             severity_text=record["level"],
             severity_number=severity_number,
             body=record["message"],
-            resource=self.logger_provider.resource,
             attributes=attributes,
         )
 
diff --git a/application_sdk/outputs/parquet.py b/application_sdk/outputs/parquet.py
index 880c2f0a4..a41e7b7d2 100644
--- a/application_sdk/outputs/parquet.py
+++ b/application_sdk/outputs/parquet.py
@@ -222,10 +222,6 @@ async def write_daft_dataframe(
             if isinstance(write_mode, str):
                 write_mode = WriteMode(write_mode)
 
-            row_count = dataframe.count_rows()
-            if row_count == 0:
-                return
-
             file_paths = []
             # Use Daft's execution context for temporary configuration
             with daft.execution_config_ctx(
@@ -239,6 +235,12 @@ async def write_daft_dataframe(
                     partition_cols=partition_cols,
                 )
                 file_paths = result.to_pydict().get("path", [])
+                        
+            # If the dataframe is empty, return 
+            # but write an empty file to the output path (necessary for ARS lookup)
+            row_count = dataframe.count_rows()
+            if row_count == 0:
+                return
 
             # Update counters
             self.chunk_count += 1