--- a
+++ b/tests/data/test_spark.py
@@ -0,0 +1,29 @@
+import edsnlp
+
+
+def test_read_write(blank_nlp, text, df_notes_pyspark):
+    # line below is just to mix params to avoid running too many tests
+    shuffle = "dataset" if blank_nlp.lang == "eds" else False
+
+    reader = edsnlp.data.from_spark(
+        df_notes_pyspark,
+        converter="omop",
+        nlp=blank_nlp,
+        shuffle=shuffle,
+    ).set_processing(backend="simple")
+    doc = list(reader)[0]
+    assert doc.text == text
+
+    blank_nlp.add_pipe("eds.matcher", config={"terms": {"douleur": ["douleurs"]}})
+    blank_nlp.add_pipe("eds.negation")
+    docs = blank_nlp.pipe(reader)
+
+    writer = edsnlp.data.to_spark(
+        docs,
+        converter="omop",
+        span_attributes=["negation"],
+        span_getter=["ents"],
+    )
+    res = writer.toPandas().to_dict(orient="records")
+    assert len(res) == 20
+    assert sum(len(r["entities"]) for r in res) == 20