From 8f7642b91078acbe69a3745f35e9b546516cd337 Mon Sep 17 00:00:00 2001
From: skrydal <piotr.skrydalewicz@acryl.io>
Date: Tue, 20 Aug 2024 13:24:16 +0200
Subject: [PATCH] fix(ingestion/tableau): Tableau field type parsing (#11202)

---
 .../ingestion/source/tableau_common.py        |  7 ++++-
 .../tests/unit/test_tableau_source.py         | 31 ++++++++++++++++++-
 2 files changed, 36 insertions(+), 2 deletions(-)

diff --git a/metadata-ingestion/src/datahub/ingestion/source/tableau_common.py b/metadata-ingestion/src/datahub/ingestion/source/tableau_common.py
index 93f2a0ef2f6a8..f3a9c4a5aa201 100644
--- a/metadata-ingestion/src/datahub/ingestion/source/tableau_common.py
+++ b/metadata-ingestion/src/datahub/ingestion/source/tableau_common.py
@@ -502,7 +502,12 @@ def get_tags_from_params(params: List[str] = []) -> GlobalTagsClass:
 
 
 def tableau_field_to_schema_field(field, ingest_tags):
-    nativeDataType = field.get("dataType", "UNKNOWN")
+    # The check here makes sure that even if 'dataType' key exists in the 'field' dictionary but has value None,
+    # it will be set as "UNKNOWN" (nativeDataType field can not be None in the SchemaField).
+    # Hence, field.get("dataType", "UNKNOWN") is not enough
+    nativeDataType = field.get("dataType")
+    if nativeDataType is None:
+        nativeDataType = "UNKNOWN"
     TypeClass = FIELD_TYPE_MAPPING.get(nativeDataType, NullTypeClass)
 
     schema_field = SchemaField(
diff --git a/metadata-ingestion/tests/unit/test_tableau_source.py b/metadata-ingestion/tests/unit/test_tableau_source.py
index f5410b161ed70..1cd0557d085f1 100644
--- a/metadata-ingestion/tests/unit/test_tableau_source.py
+++ b/metadata-ingestion/tests/unit/test_tableau_source.py
@@ -1,8 +1,37 @@
+from typing import Any, Dict
+
 import pytest
 
 import datahub.ingestion.source.tableau_constant as c
 from datahub.ingestion.source.tableau import TableauSiteSource
-from datahub.ingestion.source.tableau_common import get_filter_pages, make_filter
+from datahub.ingestion.source.tableau_common import (
+    get_filter_pages,
+    make_filter,
+    tableau_field_to_schema_field,
+)
+from datahub.metadata.com.linkedin.pegasus2avro.schema import SchemaField
+
+
+def test_tablea_source_handles_none_nativedatatype():
+    field: Dict[str, Any] = {
+        "__typename": "CalculatedField",
+        "id": "abcd",
+        "name": "Test Field",
+        "description": None,
+        "isHidden": False,
+        "folderName": None,
+        "upstreamFields": [],
+        "upstreamColumns": [],
+        "role": None,
+        "dataType": None,
+        "defaultFormat": "s",
+        "aggregation": None,
+        "formula": "a/b + d",
+    }
+    schema_field: SchemaField = tableau_field_to_schema_field(
+        field=field, ingest_tags=False
+    )
+    assert schema_field.nativeDataType == "UNKNOWN"
 
 
 def test_tableau_source_unescapes_lt():