chore(query): add combinators

Author: sundy-li

Cargo.lock

@@ -15,7 +15,7 @@ test = false
 common-base = { path = "../base" }
 
 # Crates.io dependencies
-ordered-float = "3.0.0"
+ordered-float = { git = "https://github.com/andylokandy/rust-ordered-float.git", branch = "as", features = ["serde"] }
 primitive-types = "0.11.1"
 
 [dev-dependencies]

src/common/hashtable/Cargo.toml

@@ -24,7 +24,6 @@
 #![feature(const_maybe_uninit_as_mut_ptr)]
 #![feature(const_mut_refs)]
 #![feature(generic_const_exprs)]
-
 #![allow(incomplete_features)]
 
 #[allow(dead_code)]

src/query/expression/src/lib.rs

@@ -84,8 +84,10 @@ pub const ALL_INTEGER_TYPES: &[NumberDataType; 8] = &[
     NumberDataType::Int64,
 ];
 
-pub const ALL_FLOAT_TYPES: &[NumberDataType; 2] = &[NumberDataType::Float32, NumberDataType::Float64];
-pub const ALL_NUMERICS_TYPES: &[NumberDataType; 10] = &concat_array(ALL_INTEGER_TYPES, ALL_FLOAT_TYPES);
+pub const ALL_FLOAT_TYPES: &[NumberDataType; 2] =
+    &[NumberDataType::Float32, NumberDataType::Float64];
+pub const ALL_NUMERICS_TYPES: &[NumberDataType; 10] =
+    &concat_array(ALL_INTEGER_TYPES, ALL_FLOAT_TYPES);
 
 impl DataType {
     pub fn wrap_nullable(&self) -> Self {

src/query/expression/src/types.rs

@@ -24,7 +24,6 @@ use ordered_float::OrderedFloat;
 use serde::Deserialize;
 use serde::Serialize;
 
-use crate::ColumnBuilder;
 use crate::property::Domain;
 use crate::types::ArgType;
 use crate::types::DataType;
@@ -33,6 +32,7 @@ use crate::types::ValueType;
 use crate::util::buffer_into_mut;
 use crate::values::Column;
 use crate::values::Scalar;
+use crate::ColumnBuilder;
 use crate::ScalarRef;
 
 pub type F32 = OrderedFloat<f32>;
@@ -52,11 +52,10 @@ pub type UInt64Type = NumberType<u64>;
 pub type Float32Type = NumberType<F32>;
 pub type Float64Type = NumberType<F64>;
 
-
-impl <Num: Number>  NumberType<Num> {
-   pub fn try_downcast_builder<'a>(builder: &'a mut ColumnBuilder) -> Option<&'a mut Vec<Num>> {
+impl<Num: Number> NumberType<Num> {
+    pub fn try_downcast_builder<'a>(builder: &'a mut ColumnBuilder) -> Option<&'a mut Vec<Num>> {
         match builder {
-            ColumnBuilder::Number(num) => Num::try_downcast_builder(num) ,
+            ColumnBuilder::Number(num) => Num::try_downcast_builder(num),
             _ => None,
         }
     }
@@ -831,8 +830,6 @@ fn overflow_cast<T: Number, U: Number>(src: T) -> (U, bool) {
     (dest, overflowing)
 }
 
-
-
 #[macro_export]
 macro_rules! with_number_type {
     ( | $t:tt | $($tail:tt)* ) => {
@@ -858,15 +855,27 @@ macro_rules! with_number_mapped_type {
 }
 
 pub trait Number:
-    Copy + Debug + NumCast + Default + Clone + Copy + PartialEq + Eq + PartialOrd + Ord + Sync + Send + 'static
+    Copy
+    + Debug
+    + NumCast
+    + Default
+    + Clone
+    + Copy
+    + PartialEq
+    + Eq
+    + PartialOrd
+    + Ord
+    + Sync
+    + Send
+    + 'static
 {
     const MIN: Self;
     const MAX: Self;
 
     fn data_type() -> NumberDataType;
     fn try_downcast_scalar(scalar: &NumberScalar) -> Option<Self>;
     fn try_downcast_column(col: &NumberColumn) -> Option<Buffer<Self>>;
-    //TODO
+    // TODO
     fn try_downcast_builder(col: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>>;
     fn try_downcast_domain(domain: &NumberDomain) -> Option<SimpleDomain<Self>>;
     fn upcast_scalar(scalar: Self) -> NumberScalar;
@@ -927,7 +936,7 @@ impl Number for u16 {
         col.as_u_int16().cloned()
     }
 
- fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_u_int16_mut()
     }
 
@@ -964,7 +973,7 @@ impl Number for u32 {
         col.as_u_int32().cloned()
     }
 
-     fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_u_int32_mut()
     }
 
@@ -1001,7 +1010,7 @@ impl Number for u64 {
         col.as_u_int64().cloned()
     }
 
-      fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_u_int64_mut()
     }
 
@@ -1038,7 +1047,7 @@ impl Number for i8 {
         col.as_int8().cloned()
     }
 
-   fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_int8_mut()
     }
 
@@ -1075,7 +1084,7 @@ impl Number for i16 {
         col.as_int16().cloned()
     }
 
- fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_int16_mut()
     }
 
@@ -1112,7 +1121,7 @@ impl Number for i32 {
         col.as_int32().cloned()
     }
 
-     fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_int32_mut()
     }
 
@@ -1149,15 +1158,14 @@ impl Number for i64 {
         col.as_int64().cloned()
     }
 
-   fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_int64_mut()
     }
 
     fn try_downcast_domain(domain: &NumberDomain) -> Option<SimpleDomain<Self>> {
         domain.as_int64().cloned()
     }
 
-
     fn upcast_scalar(scalar: Self) -> NumberScalar {
         NumberScalar::Int64(scalar)
     }
@@ -1187,7 +1195,7 @@ impl Number for F32 {
         col.as_float32().cloned()
     }
 
-      fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_float32_mut()
     }
 
@@ -1224,7 +1232,7 @@ impl Number for F64 {
         col.as_float64().cloned()
     }
 
-      fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
+    fn try_downcast_builder(builder: &mut NumberColumnBuilder) -> Option<&mut Vec<Self>> {
         builder.as_float64_mut()
     }
 

src/query/expression/src/types/number.rs

@@ -17,6 +17,8 @@ use std::ops::Range;
 
 use common_arrow::arrow::buffer::Buffer;
 use common_arrow::arrow::trusted_len::TrustedLen;
+use serde::Deserialize;
+use serde::Serialize;
 
 use crate::property::Domain;
 use crate::types::ArgType;
@@ -200,7 +202,7 @@ impl<'a> Iterator for StringIterator<'a> {
 
 unsafe impl<'a> TrustedLen for StringIterator<'a> {}
 
-#[derive(Debug, Clone, PartialEq, Eq)]
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
 pub struct StringColumnBuilder {
     pub data: Vec<u8>,
     pub offsets: Vec<u64>,
@@ -284,6 +286,14 @@ impl StringColumnBuilder {
         assert_eq!(self.offsets.len(), 2);
         self.data[(self.offsets[0] as usize)..(self.offsets[1] as usize)].to_vec()
     }
+
+    /// # Safety
+    pub unsafe fn index_unchecked(&self, row: usize) -> &[u8] {
+        let start = *self.offsets.get_unchecked(row) as usize;
+        let end = *self.offsets.get_unchecked(row + 1) as usize;
+        // soundness: the invariant of the struct
+        self.data.get_unchecked(start..end)
+    }
 }
 
 #[derive(Debug, Clone, PartialEq, Eq)]

src/query/expression/src/types/string.rs

@@ -92,7 +92,6 @@ pub fn deserialize_arrow_array(bytes: &[u8]) -> Option<Box<dyn Array>> {
     Some(col)
 }
 
-
 pub const fn concat_array<T, const A: usize, const B: usize>(a: &[T; A], b: &[T; B]) -> [T; A + B] {
     let mut result = std::mem::MaybeUninit::uninit();
     let dest = result.as_mut_ptr() as *mut T;