fix: allow primary key to be an empty string

mostlyai/sdk/_data/base.py

@@ -189,7 +189,7 @@ def get_context_key(self, table_name: str) -> DataIdentifier | None:
 
     def get_primary_key(self, table_name: str) -> DataIdentifier | None:
         primary_key = None
-        if self.tables[table_name].primary_key is not None:
+        if self.tables[table_name].primary_key:
             # first, check primary key on table
             primary_key = DataIdentifier(table_name, self.tables[table_name].primary_key)
         elif context_relations := self.get_child_context_relations(table_name):
@@ -260,7 +260,7 @@ def get_scp_relations(self, table: str) -> list[ContextRelation]:
 
     def _update_key_encoding_types(self) -> None:
         for tbl_name, tbl_table in self.tables.items():
-            if tbl_table.primary_key is not None:
+            if tbl_table.primary_key:
                 if tbl_table.primary_key in tbl_table.encoding_types:
                     del tbl_table.encoding_types[tbl_table.primary_key]
             for rel in self.get_relations_to_table(tbl_name):

mostlyai/sdk/_data/pull_utils.py

@@ -448,7 +448,7 @@ def fetch_table_data(
     )
     n_fetched_rows = 0
     for idx, chunk_df in enumerate(iterator):
-        if deduplicate_pks and primary_key is not None and primary_key.column in chunk_df.columns:
+        if deduplicate_pks and primary_key and primary_key.column in chunk_df.columns:
             # consider only the first occurrence of each primary key
             if not chunk_df[primary_key.column].is_unique:
                 drop_idx = chunk_df[primary_key.column].duplicated()
@@ -568,7 +568,7 @@ def fetch_target_table(
             key_fraction_df[FRACTION] = keys[MAX_TGT_ROWS_PER_CTX_KEY] / schema.tables[tgt].row_count
             key_fraction_df = key_fraction_df.rename(columns={tgt_context_key.ref_name(): tgt_context_key.column})
             key_fraction_df = key_fraction_df.drop(columns=[MAX_TGT_ROWS_PER_CTX_KEY])
-    elif tgt_primary_key is not None:
+    elif tgt_primary_key:
         # flat setup with primary key
         # sample target table by target primary key
         # use provided keys to filter and order the target table (max_sample_size=None, do_shuffle=False)

mostlyai/sdk/domain.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 
 # generated by datamodel-codegen:
-#   timestamp: 2025-06-26T08:04:20+00:00
+#   timestamp: 2025-07-09T15:31:45+00:00
 
 from __future__ import annotations
 
@@ -2735,7 +2735,7 @@ def validate_keys_exists_in_columns(cls, values):
         if values.columns:
             column_names = {col.name for col in values.columns}
             pk = values.primary_key
-            if pk is not None and pk not in column_names:
+            if pk and pk not in column_names:
                 raise ValueError(f"Primary key column '{pk}' does not exist in the table's columns.")
             for fk in values.foreign_keys or []:
                 if fk.column not in column_names:
@@ -2747,7 +2747,7 @@ def validate_keys_exists_in_columns(cls, values):
     def validate_pk_and_fks_are_not_overlapping(cls, values):
         primary_key = values.primary_key
         foreign_keys = [fk.column for fk in values.foreign_keys or []]
-        if primary_key is not None and primary_key in foreign_keys:
+        if primary_key and primary_key in foreign_keys:
             raise ValueError(f"Column '{primary_key}' is both a primary key and a foreign key.")
         return values
 

tools/custom_template/pydantic_v2/BaseModel.jinja2

@@ -786,7 +786,7 @@ class {{ class_name }}({{ base_class }}):{% if comment is defined %}  # {{ comme
         if values.columns:
             column_names = {col.name for col in values.columns}
             pk = values.primary_key
-            if pk is not None and pk not in column_names:
+            if pk and pk not in column_names:
                 raise ValueError(f"Primary key column '{pk}' does not exist in the table's columns.")
             for fk in values.foreign_keys or []:
                 if fk.column not in column_names:
@@ -798,7 +798,7 @@ class {{ class_name }}({{ base_class }}):{% if comment is defined %}  # {{ comme
     def validate_pk_and_fks_are_not_overlapping(cls, values):
         primary_key = values.primary_key
         foreign_keys = [fk.column for fk in values.foreign_keys or []]
-        if primary_key is not None and primary_key in foreign_keys:
+        if primary_key and primary_key in foreign_keys:
             raise ValueError(f"Column '{primary_key}' is both a primary key and a foreign key.")
         return values
 {%- endif %}{%- if class_name == "SyntheticTableConfiguration" %}

tools/model.py

@@ -600,7 +600,7 @@ def validate_keys_exists_in_columns(cls, values):
         if values.columns:
             column_names = {col.name for col in values.columns}
             pk = values.primary_key
-            if pk is not None and pk not in column_names:
+            if pk and pk not in column_names:
                 raise ValueError(f"Primary key column '{pk}' does not exist in the table's columns.")
             for fk in values.foreign_keys or []:
                 if fk.column not in column_names:
@@ -612,7 +612,7 @@ def validate_keys_exists_in_columns(cls, values):
     def validate_pk_and_fks_are_not_overlapping(cls, values):
         primary_key = values.primary_key
         foreign_keys = [fk.column for fk in values.foreign_keys or []]
-        if primary_key is not None and primary_key in foreign_keys:
+        if primary_key and primary_key in foreign_keys:
             raise ValueError(f"Column '{primary_key}' is both a primary key and a foreign key.")
         return values