Minor improvements

examples/parquet_read_record.rs

@@ -11,7 +11,7 @@ fn main() -> Result<()> {
     let file_path = &args[1];
 
     let reader = File::open(file_path)?;
-    let reader = read::RecordReader::try_new(reader, None, None, None, None)?;
+    let reader = read::FileReader::try_new(reader, None, None, None, None)?;
 
     let start = SystemTime::now();
     for maybe_chunk in reader {

src/io/parquet/read/binary/utils.rs

@@ -17,11 +17,6 @@ impl<O: Offset> Pushable<O> for Offsets<O> {
         self.0.len() - 1
     }
 
-    #[inline]
-    fn reserve(&mut self, additional: usize) {
-        self.0.reserve(additional)
-    }
-
     #[inline]
     fn push(&mut self, value: O) {
         self.0.push(value)
@@ -76,11 +71,6 @@ impl<'a, O: Offset> Pushable<&'a [u8]> for Binary<O> {
         self.len()
     }
 
-    #[inline]
-    fn reserve(&mut self, additional: usize) {
-        self.offsets.reserve(additional)
-    }
-
     #[inline]
     fn push_null(&mut self) {
         self.push(&[])

src/io/parquet/read/fixed_size_binary/utils.rs

@@ -30,11 +30,6 @@ impl FixedSizeBinary {
 }
 
 impl<'a> Pushable<&'a [u8]> for FixedSizeBinary {
-    #[inline]
-    fn reserve(&mut self, additional: usize) {
-        self.values.reserve(additional * self.size)
-    }
-
     #[inline]
     fn push(&mut self, value: &[u8]) {
         debug_assert_eq!(value.len(), self.size);

src/io/parquet/read/utils.rs

@@ -70,7 +70,7 @@ pub fn split_buffer<'a>(
 
 /// A private trait representing structs that can receive elements.
 pub(super) trait Pushable<T>: Sized {
-    fn reserve(&mut self, additional: usize);
+    //fn reserve(&mut self, additional: usize);
     fn push(&mut self, value: T);
     fn len(&self) -> usize;
     fn push_null(&mut self);
@@ -83,11 +83,6 @@ impl Pushable<bool> for MutableBitmap {
         self.len()
     }
 
-    #[inline]
-    fn reserve(&mut self, additional: usize) {
-        self.reserve(additional)
-    }
-
     #[inline]
     fn push(&mut self, value: bool) {
         self.push(value)
@@ -110,11 +105,6 @@ impl<A: Copy + Default> Pushable<A> for Vec<A> {
         self.len()
     }
 
-    #[inline]
-    fn reserve(&mut self, additional: usize) {
-        self.reserve(additional)
-    }
-
     #[inline]
     fn push_null(&mut self) {
         self.push(A::default())
@@ -163,7 +153,6 @@ impl<'a> OptionalPageValidity<'a> {
 }
 
 /// Extends a [`Pushable`] from an iterator of non-null values and an hybrid-rle decoder
-#[inline]
 pub(super) fn extend_from_decoder<'a, T: Default, P: Pushable<T>, I: Iterator<Item = T>>(
     validity: &mut MutableBitmap,
     page_validity: &mut OptionalPageValidity<'a>,
@@ -240,31 +229,6 @@ pub(super) fn extend_from_decoder<'a, T: Default, P: Pushable<T>, I: Iterator<It
     }
 }
 
-/*
-pub(super) fn read_dict_optional<K>(
-    validity_buffer: &[u8],
-    indices_buffer: &[u8],
-    additional: usize,
-    indices: &mut Vec<K>,
-    validity: &mut MutableBitmap,
-) where
-    K: DictionaryKey,
-{
-    // SPEC: Data page format: the bit width used to encode the entry ids stored as 1 byte (max bit width = 32),
-    // SPEC: followed by the values encoded using RLE/Bit packed described above (with the given bit width).
-    let bit_width = indices_buffer[0];
-    let indices_buffer = &indices_buffer[1..];
-
-    let new_indices =
-        hybrid_rle::HybridRleDecoder::new(indices_buffer, bit_width as u32, additional);
-    let indices_iter = new_indices.map(|x| K::from_u32(x).unwrap());
-
-    let mut page_validity = OptionalPageValidity::new(validity_buffer, additional);
-
-    extend_from_decoder(validity, &mut page_validity, None, indices, indices_iter)
-}
- */
-
 /// The state of a partially deserialized page
 pub(super) trait PageState<'a> {
     fn len(&self) -> usize;

tests/it/io/parquet/mod.rs

@@ -677,7 +677,7 @@ fn test_file(version: &str, file_name: &str) -> Result<()> {
     // empty batches are not written/read from parquet and can be ignored
     let batches = batches
         .into_iter()
-        .filter(|x| x.len() > 0)
+        .filter(|x| !x.is_empty())
         .collect::<Vec<_>>();
 
     let data = integration_write(&schema, &batches)?;