Merge pull request #788 from atlanhq/DEVX-329

DEVX-329 | Separate existence checks between keyword and text fields

Author: Aryamanz29

pyatlan/model/fields/atlan_fields.py

@@ -29,6 +29,25 @@
 from pyatlan.utils import ComparisonCategory, is_comparable_type
 
 
+class AtlanSearchableFieldType(str, Enum):
+    """
+    Enum to specify which field index to use when checking for existence of a value
+    in a KeywordTextField. This is particularly useful for the has_any_value() method.
+    """
+
+    KEYWORD = "KEYWORD"
+    """
+    Use the keyword field index (e.g., userDescription.keyword).
+    Note: Elasticsearch does not index .keyword for text exceeding 5K characters.
+    """
+
+    TEXT = "TEXT"
+    """
+    Use the text field index (e.g., userDescription.text or userDescription).
+    This should be used when field values may exceed 5K characters.
+    """
+
+
 class AtlanField(ABC):
     """
     Base enumeration of all attributes that exist in Atlan, so you do not have to remember their
@@ -75,13 +94,28 @@ def __init__(self, atlan_field_name: StrictStr, elastic_field_name: StrictStr):
     def internal_field_name(self):
         return self.atlan_field_name
 
-    def has_any_value(self) -> Query:
+    def has_any_value(
+        self,
+        field_type: Optional[AtlanSearchableFieldType] = None,
+    ) -> Query:
         """
         Returns a query that will only match assets that have some non-null, non-empty value
         (no matter what actual value) for the field.
 
+        Note: When text exceeds a particular length (5K characters), the keyword field on an
+        attribute can be empty while the text field on the same attribute is populated.
+        For KeywordTextField types, use field_type=AtlanSearchableFieldType.TEXT to check
+        the text field instead. For other field types, this parameter is ignored.
+
+        :param field_type: optional field type to check for existence (KEYWORD or TEXT).
+                          Only applicable for KeywordTextField types.
+                          Defaults to KEYWORD (None) for backwards compatibility.
         :returns: a query that will only match assets that have some non-null, non-empty value for the field
         """
+        if field_type == AtlanSearchableFieldType.TEXT and hasattr(
+            self, "_text_field_name"
+        ):
+            return Exists(field=self._text_field_name)
         return Exists(field=self.elastic_field_name)
 
     def order(self, order: SortOrder = SortOrder.ASCENDING) -> SortItem:

tests/unit/model/fields/atlan_fields_test.py

@@ -4,6 +4,7 @@
 
 from pyatlan.model.enums import SortOrder
 from pyatlan.model.fields.atlan_fields import (
+    AtlanSearchableFieldType,
     InternalKeywordTextField,
     KeywordField,
     KeywordTextField,
@@ -90,6 +91,31 @@ def test_atlan_field_name(self, sut: KeywordTextField):
     def test_keyword_field_name(self, sut: KeywordTextField):
         assert sut.keyword_field_name == KEYWORD_FIELD_NAME
 
+    def test_has_any_value_default_uses_keyword_field(self, sut: KeywordTextField):
+        """Test that has_any_value() with no arguments uses keyword field for backwards compatibility."""
+        exists = sut.has_any_value()
+
+        assert isinstance(exists, Exists)
+        assert exists.field == KEYWORD_FIELD_NAME
+
+    def test_has_any_value_with_keyword_field_type(self, sut: KeywordTextField):
+        """Test that has_any_value() with KEYWORD field_type uses keyword field."""
+        exists = sut.has_any_value(field_type=AtlanSearchableFieldType.KEYWORD)
+
+        assert isinstance(exists, Exists)
+        assert exists.field == KEYWORD_FIELD_NAME
+
+    def test_has_any_value_with_text_field_type(self, sut: KeywordTextField):
+        """Test that has_any_value() with TEXT field_type uses text field.
+
+        This is useful when field values may exceed 5K characters, as Elasticsearch
+        does not index .keyword for text exceeding 5K characters.
+        """
+        exists = sut.has_any_value(field_type=AtlanSearchableFieldType.TEXT)
+
+        assert isinstance(exists, Exists)
+        assert exists.field == TEXT_FIELD_NAME
+
 
 class TestInternalKeywordTextField:
     @pytest.fixture()