restrict arrays as parents of other arrays

Author: K-Meech

src/zarr/core/array.py

@@ -119,6 +119,7 @@
 from zarr.core.sync import sync
 from zarr.errors import (
     ArrayNotFoundError,
+    ContainsArrayError,
     MetadataValidationError,
     ZarrDeprecationWarning,
     ZarrUserWarning,
@@ -1496,14 +1497,23 @@ async def _save_metadata(self, metadata: ArrayMetadata, ensure_parents: bool = F
         Asynchronously save the array metadata.
         """
         to_save = metadata.to_buffer_dict(cpu_buffer_prototype)
-        awaitables = [set_or_delete(self.store_path / key, value) for key, value in to_save.items()]
+        set_awaitables = [
+            set_or_delete(self.store_path / key, value) for key, value in to_save.items()
+        ]
 
         if ensure_parents:
             # To enable zarr.create(store, path="a/b/c"), we need to create all the intermediate groups.
             parents = _build_parents(self)
+            ensure_array_awaitables = []
 
             for parent in parents:
-                awaitables.extend(
+                # Error if an array already exists at any parent location. Only groups can have child nodes.
+                ensure_array_awaitables.append(
+                    ensure_no_existing_node(
+                        parent.store_path, metadata.zarr_format, node_type="array"
+                    )
+                )
+                set_awaitables.extend(
                     [
                         (parent.store_path / key).set_if_not_exists(value)
                         for key, value in parent.metadata.to_buffer_dict(
@@ -1512,7 +1522,16 @@ async def _save_metadata(self, metadata: ArrayMetadata, ensure_parents: bool = F
                     ]
                 )
 
-        await gather(*awaitables)
+            # Checks for parent arrays must happen first, before any metadata is modified
+            try:
+                await gather(*ensure_array_awaitables)
+            except ContainsArrayError as e:
+                set_awaitables = []  # clear awaitables to avoid printed RuntimeWarning: coroutine was never awaited
+                raise ValueError(
+                    f"A parent of {self.store_path} is an array - only groups may have child nodes."
+                ) from e
+
+        await gather(*set_awaitables)
 
     async def _set_selection(
         self,

src/zarr/storage/_common.py

@@ -400,7 +400,11 @@ def _is_fsspec_uri(uri: str) -> bool:
     return "://" in uri or ("::" in uri and "local://" not in uri)
 
 
-async def ensure_no_existing_node(store_path: StorePath, zarr_format: ZarrFormat) -> None:
+async def ensure_no_existing_node(
+    store_path: StorePath,
+    zarr_format: ZarrFormat,
+    node_type: Literal["array", "group"] | None = None,
+) -> None:
     """
     Check if a store_path is safe for array / group creation.
     Returns `None` or raises an exception.
@@ -411,6 +415,8 @@ async def ensure_no_existing_node(store_path: StorePath, zarr_format: ZarrFormat
         The storage location to check.
     zarr_format : ZarrFormat
         The Zarr format to check.
+    node_type : str | None, optional
+        Raise an error if an "array", or "group" exists. By default (when None), raises an error for either.
 
     Raises
     ------
@@ -421,13 +427,13 @@ async def ensure_no_existing_node(store_path: StorePath, zarr_format: ZarrFormat
     elif zarr_format == 3:
         extant_node = await _contains_node_v3(store_path)
 
-    if extant_node == "array":
+    if extant_node == "array" and node_type != "group":
         raise ContainsArrayError(store_path.store, store_path.path)
-    elif extant_node == "group":
+    elif extant_node == "group" and node_type != "array":
         raise ContainsGroupError(store_path.store, store_path.path)
     elif extant_node == "nothing":
         return
-    msg = f"Invalid value for extant_node: {extant_node}"  # type: ignore[unreachable]
+    msg = f"Invalid value for extant_node: {extant_node}"
     raise ValueError(msg)