diff --git a/packages/vaex-core/vaex/__init__.py b/packages/vaex-core/vaex/__init__.py
index 50bbb0b921..673e179d18 100644
--- a/packages/vaex-core/vaex/__init__.py
+++ b/packages/vaex-core/vaex/__init__.py
@@ -379,7 +379,7 @@ def from_arrow_table(table) -> vaex.dataframe.DataFrame:
 
 
 def from_arrow_dataset(arrow_dataset) -> vaex.dataframe.DataFrame:
-    '''Create a DataFrame from an Apache Arrow dataset'''
+    '''Create a DataFrame from an Apache Arrow dataset.'''
     import vaex.arrow.dataset
     return from_dataset(vaex.arrow.dataset.DatasetArrow(arrow_dataset))
 
diff --git a/packages/vaex-core/vaex/arrow/dataset.py b/packages/vaex-core/vaex/arrow/dataset.py
index 1c08c130f6..57a6ee482c 100644
--- a/packages/vaex-core/vaex/arrow/dataset.py
+++ b/packages/vaex-core/vaex/arrow/dataset.py
@@ -250,6 +250,39 @@ def __getstate__(self):
         return state
 
 
+
+class DatasetArrow(DatasetArrowBase):
+    snake_name = "arrow-dataset"
+    def __init__(self, ds, max_rows_read=1024**2*10):
+        self._arrow_ds = ds
+        super().__init__(max_rows_read=max_rows_read)
+
+    @property
+    def _fingerprint(self):
+        return self._id
+
+    def hashed(self):
+        raise NotImplementedError
+
+    def _create_columns(self):
+        super()._create_columns()
+        # self._ids = frozendict({name: vaex.cache.fingerprint(self._fingerprint, name) for name in self._columns})
+        self._ids = frozendict()
+
+    def _create_dataset(self):
+        self._partitions = defaultdict(list) # path -> list (which will be an arrow array later on)
+        self._partition_keys = defaultdict(dict)  # path -> key -> int/index
+
+        for fragment in self._arrow_ds.get_fragments():
+            keys = pa.dataset._get_partition_keys(fragment.partition_expression)
+            for name, value in keys.items():
+                if value not in self._partitions[name]:
+                    self._partitions[name].append(value)
+                self._partition_keys[fragment.path][name] = self._partitions[name].index(value)
+        self._partitions = {name: pa.array(values) for name, values in self._partitions.items()}
+
+
+
 class DatasetArrowFileBase(vaex.dataset.Dataset):
     def __init__(self, path, fs_options, fs=None):
         super().__init__()
diff --git a/tests/from_test.py b/tests/from_test.py
index 6319f8dfd1..2562b0004e 100644
--- a/tests/from_test.py
+++ b/tests/from_test.py
@@ -1,4 +1,9 @@
+import pytest
 import vaex
+from pathlib import Path
+
+HERE = Path(__file__).parent
+
 
 def test_from_records():
     df = vaex.from_records([
@@ -26,3 +31,12 @@ def test_from_records():
     ], array_type="numpy")
     assert df.a.tolist() == [[1, 1], [11, 12], [13, 14]]
     assert df.a.shape == (3, 2)
+
+
+def test_from_arrow_dataset():
+    import pyarrow.dataset
+    path = HERE / 'data' / 'sample_arrow_dict.parquet'
+    ds = pyarrow.dataset.dataset(path)
+    df = vaex.from_arrow_dataset(ds)
+    assert df.col1.sum() == 45
+    assert df.fingerprint() == df.fingerprint()