feat: get the delta table row count based on the table history

python/deltalake/table.py

@@ -528,6 +528,29 @@ def _backwards_enumerate(
             history.append(commit)
         return history
 
+    def count(self) -> int:
+        """
+        Get the approximate row count based on file statistics added to the Delta table.
+
+        This requires that add actions have been added to the Delta table with
+        per-file statistics enabled. Because this is an optional field this
+        "count" will be less than or equal to the true row count of the table.
+        In order to get an exact number of rows a full table scan must happen
+
+        Returns:
+            The approximate number of rows for this specific table
+        """
+        total_rows = 0
+
+        for value in self.get_add_actions().column("num_records").to_pylist():
+            # Add action file statistics are optional and so while most modern
+            # tables are _likely_ to have this information it is not
+            # guaranteed.
+            if value is not None:
+                total_rows += value
+
+        return total_rows
+
     def vacuum(
         self,
         retention_hours: int | None = None,

python/tests/test_table_read.py

@@ -57,6 +57,12 @@ def test_read_simple_table_to_dict():
     ).read_all()["id"].to_pylist() == [5, 7, 9]
 
 
+def test_table_count():
+    table_path = "../crates/test/tests/data/COVID-19_NYT"
+    dt = DeltaTable(table_path)
+    assert dt.count() == 1
+
+
 class _SerializableException(BaseException):
     pass