diff --git a/examples/parquet_read.rs b/examples/parquet_read.rs
index 4fae152bfa3..4cf14fe82ef 100644
--- a/examples/parquet_read.rs
+++ b/examples/parquet_read.rs
@@ -11,7 +11,9 @@ fn main() -> Result<()> {
     let file_path = &args[1];
 
     let reader = File::open(file_path)?;
-    let reader = read::FileReader::try_new(reader, None, None, None, None)?;
+    let reader = read::FileReader::try_new(reader, Some(&[8]), None, None, None)?;
+
+    println!("{:#?}", reader.metadata());
 
     let start = SystemTime::now();
     for maybe_chunk in reader {
diff --git a/examples/parquet_read_parallel/Cargo.toml b/examples/parquet_read_parallel/Cargo.toml
index 9a102404d70..46f10ff2047 100644
--- a/examples/parquet_read_parallel/Cargo.toml
+++ b/examples/parquet_read_parallel/Cargo.toml
@@ -6,3 +6,5 @@ edition = "2021"
 [dependencies]
 arrow2 = { path = "../../", default-features = false, features = ["io_parquet", "io_parquet_compression"] }
 rayon = { version = "1", default-features = false }
+log = "0.4"
+chrono = { version = "0.4", default_features = false, features = ["std", "clock"] }
diff --git a/examples/parquet_read_parallel/src/main.rs b/examples/parquet_read_parallel/src/main.rs
index 4e1a60b401f..3b6eb243bc8 100644
--- a/examples/parquet_read_parallel/src/main.rs
+++ b/examples/parquet_read_parallel/src/main.rs
@@ -4,46 +4,67 @@ use std::io::BufReader;
 use std::sync::Arc;
 use std::time::SystemTime;
 
+use log::trace;
 use rayon::prelude::*;
 
-use arrow2::{array::Array, chunk::Chunk, error::Result, io::parquet::read};
+use arrow2::{
+    array::Array,
+    chunk::Chunk,
+    error::Result,
+    io::parquet::read::{self, ArrayIter},
+};
 
-fn parallel_read(path: &str, row_group: usize) -> Result<Chunk<Arc<dyn Array>>> {
+mod logger;
+
+/// # Panic
+/// If the iterators are empty
+fn deserialize_parallel(columns: &mut [ArrayIter<'static>]) -> Result<Chunk<Arc<dyn Array>>> {
+    // CPU-bounded
+    let columns = columns
+        .par_iter_mut()
+        .map(|iter| iter.next().transpose())
+        .collect::<Result<Vec<_>>>()?;
+
+    Chunk::try_new(columns.into_iter().map(|x| x.unwrap()).collect())
+}
+
+fn parallel_read(path: &str, row_group: usize) -> Result<()> {
     let mut file = BufReader::new(File::open(path)?);
     let metadata = read::read_metadata(&mut file)?;
     let schema = read::infer_schema(&metadata)?;
 
-    // read (IO-bounded) all columns into memory (use a subset of the fields to project)
-    let columns = read::read_columns(
-        &mut file,
-        &metadata.row_groups[row_group],
-        schema.fields,
-        None,
-    )?;
+    let row_group = &metadata.row_groups[row_group];
 
-    // CPU-bounded
-    let columns = columns
-        .into_par_iter()
-        .map(|mut iter| {
-            // when chunk_size != None, `iter` must be iterated multiple times to get all the chunks
-            // see the implementation of `arrow2::io::parquet::read::RowGroupDeserializer::next`
-            // to see how this can be done.
-            iter.next().unwrap()
-        })
-        .collect::<Result<Vec<_>>>()?;
+    let chunk_size = 1024 * 8;
 
-    Chunk::try_new(columns)
+    // read (IO-bounded) all columns into memory (use a subset of the fields to project)
+    let mut columns =
+        read::read_columns_many(&mut file, row_group, schema.fields, Some(chunk_size))?;
+
+    // deserialize (CPU-bounded) to arrow
+    let mut num_rows = row_group.num_rows();
+    while num_rows > 0 {
+        num_rows = num_rows.saturating_sub(chunk_size);
+        trace!("[parquet/deserialize][start]");
+        let chunk = deserialize_parallel(&mut columns)?;
+        trace!("[parquet/deserialize][end][{}]", chunk.len());
+        assert!(!chunk.is_empty());
+    }
+    Ok(())
 }
 
 fn main() -> Result<()> {
+    log::set_logger(&logger::LOGGER)
+        .map(|()| log::set_max_level(log::LevelFilter::Trace))
+        .unwrap();
+
     use std::env;
     let args: Vec<String> = env::args().collect();
     let file_path = &args[1];
     let row_group = args[2].parse::<usize>().unwrap();
 
     let start = SystemTime::now();
-    let batch = parallel_read(file_path, row_group)?;
-    assert!(!batch.is_empty());
+    parallel_read(file_path, row_group)?;
     println!("took: {} ms", start.elapsed().unwrap().as_millis());
 
     Ok(())