add buffered data_pages to parquet column writer total bytes estimation

parquet/src/column/writer/mod.rs

@@ -555,7 +555,12 @@ impl<'a, E: ColumnValueEncoder> GenericColumnWriter<'a, E> {
     /// of the current memory usage and not the final anticipated encoded size.
     #[cfg(feature = "arrow")]
     pub(crate) fn memory_size(&self) -> usize {
-        self.column_metrics.total_bytes_written as usize + self.encoder.estimated_memory_size()
+        self.data_pages
+            .iter()
+            .map(|page| page.data().len())
+            .sum::<usize>()
+            + self.column_metrics.total_bytes_written as usize
+            + self.encoder.estimated_memory_size()
     }
 
     /// Returns total number of bytes written by this column writer so far.
@@ -3422,6 +3427,25 @@ mod tests {
         assert!(stats.max_bytes_opt().is_none());
     }
 
+    #[test]
+    fn test_column_writer_memory_size() {
+        let page_writer = get_test_page_writer();
+        let props = Default::default();
+        let mut writer = get_test_column_writer::<Int32Type>(page_writer, 0, 0, props);
+        assert_eq!(writer.memory_size(), 256);
+
+        writer.write_batch(&[1, 2, 3, 4], None, None).unwrap();
+        writer.add_data_page().unwrap();
+        let size_with_one_page = writer.memory_size();
+        assert_eq!(size_with_one_page, 256 + 20);
+
+        writer.write_batch(&[5, 6, 7, 8], None, None).unwrap();
+        writer.add_data_page().unwrap();
+        let size_with_two_pages = writer.memory_size();
+        // different pages have different compressed lengths
+        assert_eq!(size_with_two_pages, 256 + 20 + 21);
+    }
+
     fn write_multiple_pages<T: DataType>(
         column_descr: &Arc<ColumnDescriptor>,
         pages: &[&[Option<T::T>]],