fix issue with bytes decoding

Author: alexbourret

python-connectors/api-connect_dataset/connector.py

@@ -4,7 +4,8 @@
 from rest_api_client import RestAPIClient
 from dku_utils import (
     get_dku_key_values, get_endpoint_parameters,
-    parse_keys_for_json, get_value_from_path, get_secure_credentials, decode_csv_data
+    parse_keys_for_json, get_value_from_path, get_secure_credentials,
+    decode_csv_data, decode_bytes
 )
 from dku_constants import DKUConstants
 import json
@@ -65,7 +66,7 @@ def generate_rows(self, dataset_schema=None, dataset_partitioning=None,
                 else:
                     record_count += 1
                     yield {
-                        DKUConstants.API_RESPONSE_KEY: "{}".format(data.decode())
+                        DKUConstants.API_RESPONSE_KEY: "{}".format(decode_bytes(data))
                     }
             if is_records_limit and record_count >= records_limit:
                 break

python-lib/dku_utils.py

@@ -146,8 +146,7 @@ def decode_csv_data(data):
     import csv
     import io
     json_data = None
-    if isinstance(data, bytes):
-        data = data.decode("utf-8")
+    data = decode_bytes(data)
     try:
         reader = csv.DictReader(io.StringIO(data))
         json_data = list(reader)
@@ -157,8 +156,14 @@ def decode_csv_data(data):
     return json_data
 
 
-def de_NaN(input):
-    if isinstance(input, float):
-        if math.isnan(input):
+def de_NaN(cell_content):
+    if isinstance(cell_content, float):
+        if math.isnan(cell_content):
             return ''
-    return input
+    return cell_content
+
+
+def decode_bytes(content):
+    if isinstance(content, bytes):
+        content = content.decode()
+    return content