diff --git a/parquet/src/column/writer/mod.rs b/parquet/src/column/writer/mod.rs
index 972b02d017d..d0160dbcafc 100644
--- a/parquet/src/column/writer/mod.rs
+++ b/parquet/src/column/writer/mod.rs
@@ -156,7 +156,6 @@ type ColumnCloseResult = (
 /// Typed column writer for a primitive column.
 pub type ColumnWriterImpl<'a, T> = GenericColumnWriter<'a, ColumnValueEncoderImpl<T>>;
 
-#[doc(hidden)]
 pub struct GenericColumnWriter<'a, E: ColumnValueEncoder> {
     // Column writer properties
     descr: ColumnDescPtr,
@@ -188,6 +187,9 @@ pub struct GenericColumnWriter<'a, E: ColumnValueEncoder> {
     max_column_value: Option<E::T>,
     num_column_nulls: u64,
     column_distinct_count: Option<u64>,
+
+    /// The order of encodings within the generated metadata does not impact its meaning,
+    /// but we use a BTreeSet so that the output is deterministic
     encodings: BTreeSet<Encoding>,
 
     // Reused buffers
diff --git a/parquet/src/file/statistics.rs b/parquet/src/file/statistics.rs
index 7d6ca28e611..33b7796bb16 100644
--- a/parquet/src/file/statistics.rs
+++ b/parquet/src/file/statistics.rs
@@ -398,7 +398,6 @@ impl fmt::Display for Statistics {
 /// Typed implementation for [`Statistics`].
 pub type TypedStatistics<T> = ValueStatistics<<T as DataType>::T>;
 
-#[doc(hidden)]
 /// Statistics for a particular [`ParquetValueType`]
 #[derive(Clone, Eq, PartialEq)]
 pub struct ValueStatistics<T> {