Skip to content

feat: Add NaN-counts support for parquet metadata #8158

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
Xuanwo wants to merge 5 commits into
base: main
Choose a base branch
from
Draft

feat: Add NaN-counts support for parquet metadata #8158

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
415e561
feat: Add IEEE-754 NAN count support
Xuanwo Aug 15, 2025
d8c20b0
Add more updates
Xuanwo Aug 15, 2025
c720804
Merge branch 'main' into IEEE-754
Xuanwo Aug 15, 2025
ba2ed70
Add more work
Xuanwo Aug 18, 2025
513407f
format
Xuanwo Aug 18, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions parquet/benches/metadata.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,6 +70,7 @@ fn encoded_meta() -> Vec<u8> {
distinct_count: None,
max_value: Some(vec![rng.random(); 8]),
min_value: Some(vec![rng.random(); 8]),
nan_count: None,
is_max_value_exact: Some(true),
is_min_value_exact: Some(true),
};
Expand Down
2 changes: 2 additions & 0 deletions parquet/src/arrow/arrow_writer/byte_array.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -287,6 +287,7 @@ impl FallbackEncoder {
encoding,
min_value,
max_value,
nan_count: None,
variable_length_bytes,
})
}
Expand Down Expand Up @@ -409,6 +410,7 @@ impl DictEncoder {
encoding: Encoding::RLE_DICTIONARY,
min_value,
max_value,
nan_count: None,
variable_length_bytes,
}
}
Expand Down
15 changes: 15 additions & 0 deletions parquet/src/column/writer/encoder.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,7 @@ pub struct DataPageValues<T> {
pub encoding: Encoding,
pub min_value: Option<T>,
pub max_value: Option<T>,
pub nan_count: Option<u64>,
pub variable_length_bytes: Option<i64>,
}

Expand Down Expand Up @@ -131,6 +132,7 @@ pub struct ColumnValueEncoderImpl<T: DataType> {
statistics_enabled: EnabledStatistics,
min_value: Option<T::T>,
max_value: Option<T::T>,
nan_count: Option<u64>,
bloom_filter: Option<Sbbf>,
variable_length_bytes: Option<i64>,
}
Expand All @@ -148,6 +150,17 @@ impl<T: DataType> ColumnValueEncoderImpl<T> {
// INTERVAL has undefined sort order, so don't write min/max stats for it
&& self.descr.converted_type() != ConvertedType::INTERVAL
{
// Count NaN values for floating point types
if matches!(T::T::PHYSICAL_TYPE, Type::FLOAT | Type::DOUBLE)
|| (T::T::PHYSICAL_TYPE == Type::FIXED_LEN_BYTE_ARRAY
&& self.descr.logical_type() == Some(LogicalType::Float16))
{
let nan_count = slice.iter().filter(|v| is_nan(&self.descr, *v)).count() as u64;
if nan_count > 0 {
*self.nan_count.get_or_insert(0) += nan_count;
}
}

if let Some((min, max)) = self.min_max(slice, None) {
update_min(&self.descr, &min, &mut self.min_value);
update_max(&self.descr, &max, &mut self.max_value);
Expand Down Expand Up @@ -210,6 +223,7 @@ impl<T: DataType> ColumnValueEncoder for ColumnValueEncoderImpl<T> {
bloom_filter,
min_value: None,
max_value: None,
nan_count: None,
variable_length_bytes: None,
})
}
Expand Down Expand Up @@ -304,6 +318,7 @@ impl<T: DataType> ColumnValueEncoder for ColumnValueEncoderImpl<T> {
num_values: std::mem::take(&mut self.num_values),
min_value: self.min_value.take(),
max_value: self.max_value.take(),
nan_count: self.nan_count.take(),
variable_length_bytes: self.variable_length_bytes.take(),
})
}
Expand Down
128 changes: 123 additions & 5 deletions parquet/src/column/writer/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -259,6 +259,7 @@ struct ColumnMetrics<T: Default> {
min_column_value: Option<T>,
max_column_value: Option<T>,
num_column_nulls: u64,
num_column_nans: Option<u64>,
column_distinct_count: Option<u64>,
variable_length_bytes: Option<i64>,
repetition_level_histogram: Option<LevelHistogram>,
Expand Down Expand Up @@ -1003,6 +1004,10 @@ impl<'a, E: ColumnValueEncoder> GenericColumnWriter<'a, E> {

self.column_metrics.num_column_nulls += self.page_metrics.num_page_nulls;

if let Some(nan_count) = values_data.nan_count {
*self.column_metrics.num_column_nans.get_or_insert(0) += nan_count;
}

let page_statistics = match (values_data.min_value, values_data.max_value) {
(Some(min), Some(max)) => {
// Update chunk level statistics
Expand All @@ -1016,7 +1021,8 @@ impl<'a, E: ColumnValueEncoder> GenericColumnWriter<'a, E> {
None,
Some(self.page_metrics.num_page_nulls),
false,
),
)
.with_nan_count(values_data.nan_count),
)
}
_ => None,
Expand Down Expand Up @@ -1190,6 +1196,7 @@ impl<'a, E: ColumnValueEncoder> GenericColumnWriter<'a, E> {
Some(self.column_metrics.num_column_nulls),
false,
)
.with_nan_count(self.column_metrics.num_column_nans)
.with_backwards_compatible_min_max(backwards_compatible_min_max)
.into();

Expand Down Expand Up @@ -2649,6 +2656,7 @@ mod tests {
stats.max_opt().unwrap(),
&ByteArray::from(f16::ONE + f16::ONE)
);
assert_eq!(stats.nan_count_opt(), Some(1));
}

#[test]
Expand All @@ -2665,6 +2673,7 @@ mod tests {
stats.max_opt().unwrap(),
&ByteArray::from(f16::ONE + f16::ONE)
);
assert_eq!(stats.nan_count_opt(), Some(1));
}

#[test]
Expand All @@ -2678,6 +2687,7 @@ mod tests {
assert!(stats.min_bytes_opt().is_none());
assert!(stats.max_bytes_opt().is_none());
assert!(stats.is_min_max_backwards_compatible());
assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
Expand Down Expand Up @@ -2736,24 +2746,26 @@ mod tests {
fn test_float_statistics_nan_middle() {
let stats = statistics_roundtrip::<FloatType>(&[1.0, f32::NAN, 2.0]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Float(stats) = stats {
if let Statistics::Float(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &1.0);
assert_eq!(stats.max_opt().unwrap(), &2.0);
} else {
panic!("expecting Statistics::Float");
}
assert_eq!(stats.nan_count_opt(), Some(1));
}

#[test]
fn test_float_statistics_nan_start() {
let stats = statistics_roundtrip::<FloatType>(&[f32::NAN, 1.0, 2.0]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Float(stats) = stats {
if let Statistics::Float(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &1.0);
assert_eq!(stats.max_opt().unwrap(), &2.0);
} else {
panic!("expecting Statistics::Float");
}
assert_eq!(stats.nan_count_opt(), Some(1));
}

#[test]
Expand All @@ -2763,6 +2775,7 @@ mod tests {
assert!(stats.max_bytes_opt().is_none());
assert!(stats.is_min_max_backwards_compatible());
assert!(matches!(stats, Statistics::Float(_)));
assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
Expand Down Expand Up @@ -2823,24 +2836,26 @@ mod tests {
fn test_double_statistics_nan_middle() {
let stats = statistics_roundtrip::<DoubleType>(&[1.0, f64::NAN, 2.0]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Double(stats) = stats {
if let Statistics::Double(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &1.0);
assert_eq!(stats.max_opt().unwrap(), &2.0);
} else {
panic!("expecting Statistics::Double");
}
assert_eq!(stats.nan_count_opt(), Some(1));
}

#[test]
fn test_double_statistics_nan_start() {
let stats = statistics_roundtrip::<DoubleType>(&[f64::NAN, 1.0, 2.0]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Double(stats) = stats {
if let Statistics::Double(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &1.0);
assert_eq!(stats.max_opt().unwrap(), &2.0);
} else {
panic!("expecting Statistics::Double");
}
assert_eq!(stats.nan_count_opt(), Some(1));
}

#[test]
Expand All @@ -2850,6 +2865,7 @@ mod tests {
assert!(stats.max_bytes_opt().is_none());
assert!(matches!(stats, Statistics::Double(_)));
assert!(stats.is_min_max_backwards_compatible());
assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
Expand Down Expand Up @@ -2906,6 +2922,108 @@ mod tests {
}
}

#[test]
fn test_float_statistics_infinity_with_nan() {
// Test column with Infinity and NaN values
let stats =
statistics_roundtrip::<FloatType>(&[1.0, f32::INFINITY, f32::NAN, 2.0, f32::NAN]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Float(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &1.0);
assert_eq!(stats.max_opt().unwrap(), &f32::INFINITY);
} else {
panic!("expecting Statistics::Float");
}
assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
fn test_float_statistics_neg_infinity_with_nan() {
// Test column with -Infinity and NaN values
let stats = statistics_roundtrip::<FloatType>(&[
f32::NEG_INFINITY,
-1.0,
f32::NAN,
0.0,
f32::NAN,
1.0,
]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Float(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &f32::NEG_INFINITY);
assert_eq!(stats.max_opt().unwrap(), &1.0);
} else {
panic!("expecting Statistics::Float");
}
assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
fn test_float_statistics_both_infinities_with_nan() {
// Test column with both +Infinity, -Infinity and NaN values
let stats = statistics_roundtrip::<FloatType>(&[
f32::NEG_INFINITY,
f32::NAN,
0.0,
f32::INFINITY,
f32::NAN,
]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Float(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &f32::NEG_INFINITY);
assert_eq!(stats.max_opt().unwrap(), &f32::INFINITY);
} else {
panic!("expecting Statistics::Float");
}
assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
fn test_double_statistics_infinity_with_nan() {
// Test with f64 (double) type
let stats = statistics_roundtrip::<DoubleType>(&[
1.0,
f64::INFINITY,
f64::NAN,
f64::NEG_INFINITY,
f64::NAN,
2.0,
]);
assert!(stats.is_min_max_backwards_compatible());
if let Statistics::Double(stats) = &stats {
assert_eq!(stats.min_opt().unwrap(), &f64::NEG_INFINITY);
assert_eq!(stats.max_opt().unwrap(), &f64::INFINITY);
} else {
panic!("expecting Statistics::Double");
}

assert_eq!(stats.nan_count_opt(), Some(2));
}

#[test]
fn test_float16_statistics_infinity_with_nan() {
// Test Float16 with Infinity and NaN
let input = [
f16::ONE,
f16::INFINITY,
f16::NAN,
f16::NEG_INFINITY,
f16::NAN,
]
.into_iter()
.map(|s| ByteArray::from(s).into())
.collect::<Vec<_>>();

let stats = float16_statistics_roundtrip(&input);
assert!(stats.is_min_max_backwards_compatible());
assert_eq!(
stats.min_opt().unwrap(),
&ByteArray::from(f16::NEG_INFINITY)
);
assert_eq!(stats.max_opt().unwrap(), &ByteArray::from(f16::INFINITY));
assert_eq!(stats.nan_count_opt(), Some(2));
}
Copy link
Copy Markdown
Contributor

@alamb alamb Aug 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the spec also talks about the case for all nans, so that would also be a good case to check

Copy link
Copy Markdown
Member Author

@Xuanwo Xuanwo Aug 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you for the review, I will polish the tests side.


#[test]
fn test_compare_greater_byte_array_decimals() {
assert!(!compare_greater_byte_array_decimals(&[], &[],),);
Expand Down
8 changes: 4 additions & 4 deletions parquet/src/file/metadata/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1968,9 +1968,9 @@ mod tests {
.build();

#[cfg(not(feature = "encryption"))]
let base_expected_size = 2312;
let base_expected_size = 2376;
#[cfg(feature = "encryption")]
let base_expected_size = 2648;
let base_expected_size = 2712;

assert_eq!(parquet_meta.memory_size(), base_expected_size);

Expand Down Expand Up @@ -1998,9 +1998,9 @@ mod tests {
.build();

#[cfg(not(feature = "encryption"))]
let bigger_expected_size = 2816;
let bigger_expected_size = 2880;
#[cfg(feature = "encryption")]
let bigger_expected_size = 3152;
let bigger_expected_size = 3216;

// more set fields means more memory usage
assert!(bigger_expected_size > base_expected_size);
Expand Down
Loading
Loading