feat: add type system encoding to text bytes

This commit is contained in:
Jindřich Moravec 2024-01-23 21:55:43 +01:00
parent f47fd24232
commit 7773a2585e
2 changed files with 170 additions and 0 deletions

View file

@ -1,3 +1,5 @@
use std::num::{ParseFloatError, ParseIntError};
use std::str::Utf8Error;
use thiserror::Error; use thiserror::Error;
use crate::internals::row::ColumnPosition; use crate::internals::row::ColumnPosition;
use crate::schema::{ColumnName, TableName}; use crate::schema::{ColumnName, TableName};
@ -25,3 +27,18 @@ pub enum Error {
#[error("table {0} cannot be indexed on column {1}")] #[error("table {0} cannot be indexed on column {1}")]
AttemptToIndexNonIndexableColumn(TableName, ColumnName), AttemptToIndexNonIndexableColumn(TableName, ColumnName),
} }
#[derive(Debug, Error)]
pub enum TypeConversionError {
#[error("failed to decode bytes to string")]
TextDecodeFailed(#[from] Utf8Error),
#[error("failed to parse float from text")]
NumberDecodeFailed(#[from] ParseFloatError),
#[error("failed to parse int from text")]
IntDecodeFailed(#[from] ParseIntError),
#[error("uknown type with oid {oid} and size {size}")]
UnknownType {
oid: i32,
size: i32
}
}

View file

@ -1,3 +1,5 @@
use crate::error::TypeConversionError;
// ==============Types================ // ==============Types================
#[derive(Debug, Clone, Copy)] #[derive(Debug, Clone, Copy)]
pub enum DbType { pub enum DbType {
@ -39,4 +41,155 @@ impl Value {
}, },
} }
} }
pub fn type_oid(&self) -> i32 {
match self {
Self::Number(_) => 701,
Self::Indexable(val) => match val {
IndexableValue::String(_) => 25,
IndexableValue::Int(_) => 23,
IndexableValue::Uuid(_) => 2950,
},
}
}
pub fn type_size(&self) -> i32 {
match self {
Self::Number(_) => 8,
Self::Indexable(val) => match val {
IndexableValue::String(_) => -2, // null terminated string
IndexableValue::Int(_) => 8,
IndexableValue::Uuid(_) => 16,
},
}
}
pub fn as_text_bytes(&self) -> Vec<u8> {
match self {
Self::Number(n) => format!("{n}").into_bytes(),
Self::Indexable(i) => match i {
IndexableValue::String(s) => format!("{s}\0").into_bytes(),
IndexableValue::Int(i) => format!("{i}").into_bytes(),
IndexableValue::Uuid(u) => format!("{u}").into_bytes(),
},
}
}
pub fn from_text_bytes(bytes: &[u8], type_oid: i32, type_size: i32) -> Result<Value, TypeConversionError> {
match (type_oid, type_size) {
(701, 8) => {
let s = std::str::from_utf8(bytes)?;
let n = s.parse::<f64>()?;
Ok(Value::Number(n))
}
(25, -2) => {
let s = std::str::from_utf8(bytes)?;
let s = &s[..s.len() - 1]; // remove null terminator
Ok(Value::Indexable(IndexableValue::String(s.to_string())))
}
(23, 8) => {
let s = std::str::from_utf8(bytes)?;
let n = s.parse::<u64>()?;
Ok(Value::Indexable(IndexableValue::Int(n)))
}
(2950, 16) => {
let s = std::str::from_utf8(bytes)?;
let n = s.parse::<Uuid>()?;
Ok(Value::Indexable(IndexableValue::Uuid(n)))
}
(oid, size) => Err(TypeConversionError::UnknownType { oid, size }),
}
}
}
mod tests {
use crate::error::TypeConversionError::UnknownType;
use super::{Value, IndexableValue};
#[test]
fn test_encode_number() {
let value = Value::Number(123.456);
let oid = value.type_oid();
let size = value.type_size();
let bytes = value.as_text_bytes();
let from_bytes = Value::from_text_bytes(&bytes, oid, size).unwrap();
assert_eq!(value, from_bytes);
assert_eq!(oid, 701);
assert_eq!(size, 8);
}
#[test]
fn test_encode_string() {
let value = Value::Indexable(IndexableValue::String("hello".to_string()));
let oid = value.type_oid();
let size = value.type_size();
let bytes = value.as_text_bytes();
let from_bytes = Value::from_text_bytes(&bytes, oid, size).unwrap();
assert_eq!(value, from_bytes);
assert_eq!(oid, 25);
assert_eq!(size, -2);
}
#[test]
fn test_encode_string_utf8() {
let value = Value::Indexable(IndexableValue::String("#速度与激情9 早上好中国 现在我有冰激淋 我很喜欢冰激淋 但是《速度与激情9》比冰激淋 🍧🍦🍨".to_string()));
let oid = value.type_oid();
let size = value.type_size();
let bytes = value.as_text_bytes();
let from_bytes = Value::from_text_bytes(&bytes, oid, size).unwrap();
assert_eq!(value, from_bytes);
assert_eq!(oid, 25);
assert_eq!(size, -2);
}
#[test]
fn test_encode_int() {
let value = Value::Indexable(IndexableValue::Int(123));
let oid = value.type_oid();
let size = value.type_size();
let bytes = value.as_text_bytes();
let from_bytes = Value::from_text_bytes(&bytes, oid, size).unwrap();
assert_eq!(value, from_bytes);
assert_eq!(oid, 23);
assert_eq!(size, 8);
}
#[test]
fn test_encode_uuid() {
let value = Value::Indexable(IndexableValue::Uuid(123));
let oid = value.type_oid();
let size = value.type_size();
let bytes = value.as_text_bytes();
let from_bytes = Value::from_text_bytes(&bytes, oid, size).unwrap();
assert_eq!(value, from_bytes);
assert_eq!(oid, 2950);
assert_eq!(size, 16);
}
#[test]
fn test_mismatched_size() {
let value = Value::Indexable(IndexableValue::Uuid(123));
let oid = value.type_oid();
let size = 8;
let bytes = value.as_text_bytes();
let from_bytes = Value::from_text_bytes(&bytes, oid, size);
assert!(matches!(from_bytes, Err(UnknownType { oid: 2950, size: 8 })))
}
} }