fix: deduplicate ZipStore central directory on close

ZIP files are append-only: writing the same key (e.g. zarr.json) a
second time adds a new entry without removing the old one.  When an
array is resized, zarr.json is rewritten, which accumulates duplicates
in the central directory and triggers UserWarning: Duplicate name.

On close(), _dedup_central_directory() now scans filelist in reverse
and keeps only the last (most recent) entry for each filename, so the
on-disk central directory is clean.

Closes #3580

Made-with: Cursor

Author: Leo Ji

changes/3580.bugfix.md

@@ -0,0 +1 @@
+Fix `ZipStore` leaving duplicate entries in the zip central directory when array metadata is written more than once (e.g. after `resize()`). The central directory is now deduplicated on `close()`, keeping only the most recent entry for each filename.

src/zarr/storage/_zip.py

@@ -122,8 +122,30 @@ def close(self) -> None:
         # docstring inherited
         super().close()
         with self._lock:
+            self._dedup_central_directory()
             self._zf.close()
 
+    def _dedup_central_directory(self) -> None:
+        """Remove duplicate entries from the zip central directory.
+
+        When an array is resized or metadata is updated multiple times, ZipStore
+        writes a new entry for the same key each time (ZIP files are append-only).
+        This leaves duplicate filenames in the central directory, which confuses
+        many zip readers and wastes space.  Before closing, we keep only the
+        *last* (most recent) entry for every filename so that the on-disk central
+        directory is clean.
+        """
+        if not self._zf.mode in ("w", "a", "x"):
+            return
+        seen: set[str] = set()
+        deduped: list[zipfile.ZipInfo] = []
+        for info in reversed(self._zf.filelist):
+            if info.filename not in seen:
+                seen.add(info.filename)
+                deduped.append(info)
+        self._zf.filelist = list(reversed(deduped))
+        self._zf.NameToInfo = {info.filename: info for info in self._zf.filelist}
+
     async def clear(self) -> None:
         # docstring inherited
         with self._lock:

tests/test_store/test_zip.py

@@ -152,3 +152,28 @@ async def test_move(self, tmp_path: Path) -> None:
         assert destination.exists()
         assert not origin.exists()
         assert np.array_equal(array[...], np.arange(10))
+
+    def test_no_duplicate_entries_after_resize(self, tmp_path: Path) -> None:
+        # Regression test for https://github.com/zarr-developers/zarr-python/issues/3580
+        # Resizing an array (which rewrites zarr.json) used to leave duplicate
+        # filenames in the zip central directory.
+        zip_path = tmp_path / "data.zip"
+        store = ZipStore(zip_path, mode="w")
+        arr = zarr.create_array(store, shape=(5,), chunks=(5,), dtype="i4")
+        arr[:] = np.arange(5)
+
+        # Resize triggers metadata rewrite, producing a second zarr.json entry
+        arr.resize((10,))
+        arr[5:] = np.arange(5, 10)
+        store.close()
+
+        # Verify no duplicate filenames in the central directory
+        with zipfile.ZipFile(zip_path, "r") as zf:
+            names = [info.filename for info in zf.infolist()]
+        assert len(names) == len(set(names)), f"Duplicate entries found: {names}"
+
+        # Verify data integrity
+        store2 = ZipStore(zip_path, mode="r")
+        arr2 = zarr.open_array(store2)
+        assert arr2.shape == (10,)
+        assert np.array_equal(arr2[:], np.arange(10))