perf: Improve parse json performance

Author: b41sh

Cargo.toml

@@ -27,6 +27,8 @@ rust-version = "1.80"
 
 [dependencies]
 byteorder = "1.5.0"
+compact_str = "0.9.0"
+enum-as-inner = "0.6.1"
 ethnum = "1.5.1"
 fast-float2 = "0.2.3"
 itoa = "1.0"

src/core/databend/ser.rs

@@ -246,6 +246,48 @@ impl OwnedJsonb {
 }
 
 impl Number {
+    #[inline]
+    pub(crate) fn memory_size(&self) -> usize {
+        match self {
+            Self::Int64(v) => {
+                if *v == 0 {
+                    1
+                } else if *v >= i8::MIN.into() && *v <= i8::MAX.into() {
+                    2
+                } else if *v >= i16::MIN.into() && *v <= i16::MAX.into() {
+                    3
+                } else if *v >= i32::MIN.into() && *v <= i32::MAX.into() {
+                    5
+                } else {
+                    9
+                }
+            }
+            Self::UInt64(v) => {
+                if *v == 0 {
+                    1
+                } else if *v <= u8::MAX.into() {
+                    2
+                } else if *v <= u16::MAX.into() {
+                    3
+                } else if *v <= u32::MAX.into() {
+                    5
+                } else {
+                    9
+                }
+            }
+            Self::Float64(v) => {
+                if v.is_nan() || v.is_infinite() {
+                    1
+                } else {
+                    9
+                }
+            }
+            Self::Decimal64(_) => 10,
+            Self::Decimal128(_) => 18,
+            Self::Decimal256(_) => 34,
+        }
+    }
+
     #[inline]
     pub(crate) fn compact_encode<W: Write>(&self, mut writer: W) -> Result<usize> {
         match self {

src/core/databend/util.rs

@@ -35,13 +35,14 @@ pub enum ParseErrorCode {
     InvalidLoneLeadingSurrogateInHexEscape(u16),
     InvalidSurrogateInHexEscape(u16),
     UnexpectedEndOfHexEscape,
+    ObjectDuplicateKey(String),
 }
 
 pub type Result<T> = std::result::Result<T, Error>;
 
 impl Display for ParseErrorCode {
     fn fmt(&self, f: &mut core::fmt::Formatter) -> core::fmt::Result {
-        match *self {
+        match self {
             ParseErrorCode::InvalidEOF => f.write_str("EOF while parsing a value"),
             ParseErrorCode::InvalidNumberValue => f.write_str("invalid number"),
             ParseErrorCode::InvalidStringValue => f.write_str("invalid string"),
@@ -68,6 +69,9 @@ impl Display for ParseErrorCode {
                 write!(f, "invalid surrogate in hex escape '{:X}'", n)
             }
             ParseErrorCode::UnexpectedEndOfHexEscape => f.write_str("unexpected end of hex escape"),
+            ParseErrorCode::ObjectDuplicateKey(key) => {
+                write!(f, "duplicate object attribute \"{}\"", key)
+            }
         }
     }
 }

src/error.rs

@@ -89,7 +89,12 @@ pub use number::Number;
 pub use owned::to_owned_jsonb;
 pub use owned::OwnedJsonb;
 pub use parser::from_slice;
+pub use parser::parse_owned_jsonb;
+pub use parser::parse_owned_jsonb_standard_mode;
+pub use parser::parse_owned_jsonb_standard_mode_with_buf;
+pub use parser::parse_owned_jsonb_with_buf;
 pub use parser::parse_value;
+pub use parser::parse_value_standard_mode;
 pub use raw::from_raw_jsonb;
 pub use raw::RawJsonb;
 pub use value::*;