Fix: PowerBI .pbit parser fails on multiline DAX expressions

ingestion/src/metadata/ingestion/source/dashboard/powerbi/metadata.py

@@ -549,7 +549,9 @@ def _get_child_measures(self, table: PowerBiTable) -> List[Column]:
                 measure_type = DataType.MEASURE_VISIBLE
                 if measure.isHidden:
                     measure_type = DataType.MEASURE_HIDDEN
-                expression_text = f"Expression : {measure.expression}"
+                expression_text = (
+                    f"Expression : {measure.expression}" if measure.expression else ""
+                )
                 description_text = (
                     f"Description : {measure.description}"
                     if measure.description

ingestion/src/metadata/ingestion/source/dashboard/powerbi/models.py

@@ -12,9 +12,9 @@
 PowerBI Models
 """
 from datetime import datetime
-from typing import List, Optional
+from typing import List, Optional, Union
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, field_validator
 from typing_extensions import Annotated
 
 
@@ -135,17 +135,31 @@ class PowerBiMeasures(BaseModel):
     """
 
     name: str
-    expression: str
+    expression: Optional[Union[str, List[str]]] = None
     description: Optional[str] = None
     isHidden: Optional[bool] = False
 
+    @field_validator("expression", mode="before")
+    @classmethod
+    def normalize_expression(cls, v):
+        if isinstance(v, list):
+            return "\n".join(v)
+        return v
+
 
 class PowerBITableSource(BaseModel):
     """
     PowerBI Table Source
     """
 
-    expression: Optional[str] = None
+    expression: Optional[Union[str, List[str]]] = None
+
+    @field_validator("expression", mode="before")
+    @classmethod
+    def normalize_expression(cls, v):
+        if isinstance(v, list):
+            return "\n".join(v)
+        return v
 
 
 class PowerBiTable(BaseModel):
@@ -173,7 +187,14 @@ class TablesResponse(BaseModel):
 
 class DatasetExpression(BaseModel):
     name: str
-    expression: Optional[str] = None
+    expression: Optional[Union[str, List[str]]] = None
+
+    @field_validator("expression", mode="before")
+    @classmethod
+    def normalize_expression(cls, v):
+        if isinstance(v, list):
+            return "\n".join(v)
+        return v
 
 
 class UpstreaDataflow(BaseModel):

ingestion/tests/unit/test_powerbi_table_measures.py

@@ -4,8 +4,10 @@
 from metadata.generated.schema.type.basic import Markdown
 from metadata.ingestion.source.dashboard.powerbi.metadata import PowerbiSource
 from metadata.ingestion.source.dashboard.powerbi.models import (
+    DatasetExpression,
     PowerBiMeasures,
     PowerBiTable,
+    PowerBITableSource,
 )
 
 test_cases = {
@@ -68,6 +70,32 @@
             )
         ],
     },
+    "multiline_expression_as_list": {
+        "input": [
+            PowerBiMeasures(
+                name="multiline_measure",
+                expression=[
+                    "",
+                    "VAR PL_Net_Selection =",
+                    "    SUMX('DB P&L', [Value] * RELATED('Map Account'[Sign]))",
+                    "",
+                    "RETURN PL_Net_Selection / 1000000",
+                ],
+                description="Multiline DAX",
+                isHidden=False,
+            )
+        ],
+        "expected": [
+            Column(
+                name="multiline_measure",
+                dataType=DataType.MEASURE_VISIBLE,
+                dataTypeDisplay=DataType.MEASURE_VISIBLE,
+                description=Markdown(
+                    "Expression : \nVAR PL_Net_Selection =\n    SUMX('DB P&L', [Value] * RELATED('Map Account'[Sign]))\n\nRETURN PL_Net_Selection / 1000000\n\nDescription : Multiline DAX"
+                ),
+            )
+        ],
+    },
 }
 
 
@@ -93,3 +121,52 @@ def test_get_child_measures(test_case_name, test_case):
         assert actual_col.dataType == expected_col.dataType
         assert actual_col.dataTypeDisplay == expected_col.dataTypeDisplay
         assert actual_col.description == expected_col.description
+
+
+def test_powerbi_measures_string_expression():
+    measure = PowerBiMeasures(name="test", expression="SUM(Sales)", isHidden=False)
+    assert measure.expression == "SUM(Sales)"
+
+
+def test_powerbi_measures_list_expression():
+    measure = PowerBiMeasures(
+        name="test",
+        expression=["", "VAR x = 1", "RETURN x"],
+        isHidden=False,
+    )
+    assert measure.expression == "\nVAR x = 1\nRETURN x"
+
+
+def test_powerbi_measures_empty_list_expression():
+    measure = PowerBiMeasures(name="test", expression=[], isHidden=False)
+    assert measure.expression == ""
+
+
+def test_powerbi_table_source_string_expression():
+    source = PowerBITableSource(expression="SELECT * FROM table")
+    assert source.expression == "SELECT * FROM table"
+
+
+def test_powerbi_table_source_list_expression():
+    source = PowerBITableSource(expression=["let", "    Source = ...", "in Source"])
+    assert source.expression == "let\n    Source = ...\nin Source"
+
+
+def test_powerbi_table_source_none_expression():
+    source = PowerBITableSource(expression=None)
+    assert source.expression is None
+
+
+def test_dataset_expression_string():
+    expr = DatasetExpression(name="test", expression="SUM(Sales)")
+    assert expr.expression == "SUM(Sales)"
+
+
+def test_dataset_expression_list():
+    expr = DatasetExpression(name="test", expression=["", "VAR x = 1", "RETURN x"])
+    assert expr.expression == "\nVAR x = 1\nRETURN x"
+
+
+def test_dataset_expression_none():
+    expr = DatasetExpression(name="test", expression=None)
+    assert expr.expression is None