MINOR - Fix General Profiler Bugs (#17995)

* fix import issue

* fix: better handle None values in profiler processing

* fix: profiler errors

* chore: fix comment

* style: fix python linting failure

* fix: null byte error with the database client

---------

Co-authored-by: Chirag Madlani <12962843+chirag-madlani@users.noreply.github.com>
(cherry picked from commit 8dc6b7d282a700c2a3d4659f904a3966e16843a2)

ingestion/src/metadata/profiler/metrics/composed/null_ratio.py

@@ -50,8 +50,9 @@ class NullRatio(ComposedMetric):
         results of other Metrics
         """
 
-        count = res.get(Count.name())
-        null_count = res.get(NullCount.name())
-        if count + null_count == 0:
+        count = res.get(Count.name(), 0)
+        null_count = res.get(NullCount.name(), 0)
+        total = count + null_count
+        if total == 0:
             return None
-        return null_count / (null_count + count)
+        return null_count / total

ingestion/src/metadata/profiler/orm/functions/table_metric_computer.py

@@ -176,6 +176,8 @@ class SnowflakeTableMetricComputer(BaseTableMetricComputer):
         )
 
         rest = self._runner._session.execute(query).first()
+        if not rest:
+            return None
         if rest.rowCount is None:
             # if we don't have any row count, fallback to the base logic
             return super().compute()

ingestion/src/metadata/profiler/orm/types/custom_hex_byte_string.py

@@ -22,6 +22,7 @@ from sqlalchemy.sql.sqltypes import String, TypeDecorator
 from metadata.utils.logger import ingestion_logger
 
 logger = ingestion_logger()
+NULL_BYTE = "\x00"
 
 
 class HexByteString(TypeDecorator):
@@ -63,10 +64,22 @@ class HexByteString(TypeDecorator):
         detected_encoding = chardet.detect(bytes_value).get("encoding")
         if detected_encoding:
             try:
-                value = bytes_value.decode(encoding=detected_encoding)
-                return value
+                # Decode the bytes value with the detected encoding and replace errors with "?"
+                # if bytes cannot be decoded e.g. b"\x66\x67\x67\x9c", if detected_encoding="utf-8"
+                # will result in 'foo<6F>' (instead of failing)
+                str_value = bytes_value.decode(
+                    encoding=detected_encoding, errors="replace"
+                )
+                # Replace NULL_BYTE with empty string to avoid errors with
+                # the database client (should be O(n))
+                str_value = (
+                    str_value.replace(NULL_BYTE, "")
+                    if NULL_BYTE in str_value
+                    else str_value
+                )
+                return str_value
             except Exception as exc:
-                logger.debug("Failed to parse bytes valud as string: %s", exc)
+                logger.debug("Failed to parse bytes value as string: %s", exc)
                 logger.debug(traceback.format_exc())
 
         return value.hex()