From ae58bb96dee26d4ef346cfd29f8c7f598db00fd8 Mon Sep 17 00:00:00 2001
From: Adam Gutglick <adam@spiraldb.com>
Date: Thu, 5 Dec 2024 14:46:45 -0500
Subject: [PATCH] fix: Consistent metadata table column names for DataFusion
 stats (#1577)

Fixes #1576.
---
 vortex-array/src/stats/mod.rs                 | 30 +++++++++++--------
 .../src/persistent/statistics.rs              |  9 +++---
 2 files changed, 22 insertions(+), 17 deletions(-)

diff --git a/vortex-array/src/stats/mod.rs b/vortex-array/src/stats/mod.rs
index 517f7a57d1..afd50a4070 100644
--- a/vortex-array/src/stats/mod.rs
+++ b/vortex-array/src/stats/mod.rs
@@ -98,6 +98,22 @@ impl Stat {
             Stat::UncompressedSizeInBytes => DType::Primitive(PType::U64, NonNullable),
         }
     }
+
+    pub fn name(&self) -> &str {
+        match self {
+            Self::BitWidthFreq => "bit_width_frequency",
+            Self::TrailingZeroFreq => "trailing_zero_frequency",
+            Self::IsConstant => "is_constant",
+            Self::IsSorted => "is_sorted",
+            Self::IsStrictSorted => "is_strict_sorted",
+            Self::Max => "max",
+            Self::Min => "min",
+            Self::RunCount => "run_count",
+            Self::TrueCount => "true_count",
+            Self::NullCount => "null_count",
+            Self::UncompressedSizeInBytes => "uncompressed_size_in_bytes",
+        }
+    }
 }
 
 pub fn as_stat_bitset_bytes(stats: &[Stat]) -> Vec<u8> {
@@ -134,19 +150,7 @@ pub fn stats_from_bitset_bytes(bytes: &[u8]) -> Vec<Stat> {
 
 impl Display for Stat {
     fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
-        match self {
-            Self::BitWidthFreq => write!(f, "bit_width_frequency"),
-            Self::TrailingZeroFreq => write!(f, "trailing_zero_frequency"),
-            Self::IsConstant => write!(f, "is_constant"),
-            Self::IsSorted => write!(f, "is_sorted"),
-            Self::IsStrictSorted => write!(f, "is_strict_sorted"),
-            Self::Max => write!(f, "max"),
-            Self::Min => write!(f, "min"),
-            Self::RunCount => write!(f, "run_count"),
-            Self::TrueCount => write!(f, "true_count"),
-            Self::NullCount => write!(f, "null_count"),
-            Self::UncompressedSizeInBytes => write!(f, "uncompressed_size_in_bytes"),
-        }
+        write!(f, "{}", self.name())
     }
 }
 
diff --git a/vortex-datafusion/src/persistent/statistics.rs b/vortex-datafusion/src/persistent/statistics.rs
index e181bf6f5a..99409676db 100644
--- a/vortex-datafusion/src/persistent/statistics.rs
+++ b/vortex-datafusion/src/persistent/statistics.rs
@@ -5,6 +5,7 @@ use datafusion_common::stats::Precision;
 use datafusion_common::ColumnStatistics;
 use datafusion_expr::Accumulator;
 use vortex_array::array::StructArray;
+use vortex_array::stats::Stat;
 use vortex_array::variants::StructArrayTrait as _;
 use vortex_array::IntoCanonical;
 use vortex_error::VortexResult;
@@ -12,7 +13,7 @@ use vortex_error::VortexResult;
 pub fn array_to_col_statistics(array: &StructArray) -> VortexResult<ColumnStatistics> {
     let mut stats = ColumnStatistics::new_unknown();
 
-    if let Some(null_count_array) = array.field_by_name("null_count") {
+    if let Some(null_count_array) = array.field_by_name(Stat::NullCount.name()) {
         let array = null_count_array.into_canonical()?.into_arrow()?;
         let array = array.as_primitive::<UInt64Type>();
 
@@ -20,7 +21,7 @@ pub fn array_to_col_statistics(array: &StructArray) -> VortexResult<ColumnStatis
         stats.null_count = Precision::Exact(null_count as usize);
     }
 
-    if let Some(max_value_array) = array.field_by_name("max") {
+    if let Some(max_value_array) = array.field_by_name(Stat::Max.name()) {
         let array = max_value_array.into_canonical()?.into_arrow()?;
         let mut acc = MaxAccumulator::try_new(array.data_type())?;
         acc.update_batch(&[array])?;
@@ -29,7 +30,7 @@ pub fn array_to_col_statistics(array: &StructArray) -> VortexResult<ColumnStatis
         stats.max_value = Precision::Exact(max_val)
     }
 
-    if let Some(min_value_array) = array.field_by_name("min") {
+    if let Some(min_value_array) = array.field_by_name(Stat::Min.name()) {
         let array = min_value_array.into_canonical()?.into_arrow()?;
         let mut acc = MinAccumulator::try_new(array.data_type())?;
         acc.update_batch(&[array])?;
@@ -42,7 +43,7 @@ pub fn array_to_col_statistics(array: &StructArray) -> VortexResult<ColumnStatis
 }
 
 pub fn uncompressed_col_size(array: &StructArray) -> VortexResult<Option<u64>> {
-    match array.field_by_name("uncompressed_size") {
+    match array.field_by_name(Stat::UncompressedSizeInBytes.name()) {
         None => Ok(None),
         Some(array) => {
             let array = array.into_canonical()?.into_arrow()?;