Skip to content
This repository has been archived by the owner on Feb 18, 2024. It is now read-only.

Improved performance of decimal arithmetics #605

Merged
merged 3 commits into from
Nov 14, 2021
Merged

Improved performance of decimal arithmetics #605

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
7b41519
Improved performance of decimal addition.
jorgecarleitao Nov 14, 2021
12a2a5b
Simplified decimal arithmetics.
jorgecarleitao Nov 14, 2021
41e98d4
Removed expensive compute in hot loop.
jorgecarleitao Nov 14, 2021
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
170 changes: 56 additions & 114 deletions src/compute/arithmetics/decimal/add.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,21 +1,4 @@
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.

//! Defines the addition arithmetic kernels for Decimal `PrimitiveArrays`.
//! Defines the addition arithmetic kernels for [`PrimitiveArray`] representing decimals.
use crate::compute::arithmetics::basic::check_same_len;
use crate::{
array::{Array, PrimitiveArray},
Expand All @@ -31,12 +14,14 @@ use crate::{
error::{ArrowError, Result},
};

use super::{adjusted_precision_scale, max_value, number_digits};
use super::{adjusted_precision_scale, get_parameters, max_value, number_digits};

/// Adds two decimal primitive arrays with the same precision and scale. If the
/// precision and scale is different, then an InvalidArgumentError is returned.
/// This function panics if the added numbers result in a number larger than
/// the possible number for the selected precision.
/// Adds two decimal [`PrimitiveArray`] with the same precision and scale.
/// # Error
/// Errors if the precision and scale are different.
/// # Panic
/// This function panics iff the added numbers result in a number larger than
/// the possible number for the precision.
///
/// # Examples
/// ```
Expand All @@ -53,38 +38,22 @@ use super::{adjusted_precision_scale, max_value, number_digits};
/// assert_eq!(result, expected);
/// ```
pub fn add(lhs: &PrimitiveArray<i128>, rhs: &PrimitiveArray<i128>) -> Result<PrimitiveArray<i128>> {
// Matching on both data types from both arrays
// This match will be true only when precision and scale from both
// arrays are the same, otherwise it will return and ArrowError
match (lhs.data_type(), rhs.data_type()) {
(DataType::Decimal(lhs_p, lhs_s), DataType::Decimal(rhs_p, rhs_s)) => {
if lhs_p == rhs_p && lhs_s == rhs_s {
// Closure for the binary operation. This closure will panic if
// the sum of the values is larger than the max value possible
// for the decimal precision
let op = move |a, b| {
let res: i128 = a + b;
let (precision, _) = get_parameters(lhs.data_type(), rhs.data_type())?;

assert!(
!(res.abs() > max_value(*lhs_p)),
"Overflow in addition presented for precision {}",
lhs_p
);
let max = max_value(precision);
let op = move |a, b| {
let res: i128 = a + b;

res
};
assert!(
res.abs() <= max,
"Overflow in addition presented for precision {}",
precision
);

binary(lhs, rhs, lhs.data_type().clone(), op)
} else {
Err(ArrowError::InvalidArgumentError(
"Arrays must have the same precision and scale".to_string(),
))
}
}
_ => Err(ArrowError::InvalidArgumentError(
"Incorrect data type for the array".to_string(),
)),
}
res
};

binary(lhs, rhs, lhs.data_type().clone(), op)
}

/// Saturated addition of two decimal primitive arrays with the same precision
Expand All @@ -111,40 +80,24 @@ pub fn saturating_add(
lhs: &PrimitiveArray<i128>,
rhs: &PrimitiveArray<i128>,
) -> Result<PrimitiveArray<i128>> {
// Matching on both data types from both arrays. This match will be true
// only when precision and scale from both arrays are the same, otherwise
// it will return and ArrowError
match (lhs.data_type(), rhs.data_type()) {
(DataType::Decimal(lhs_p, lhs_s), DataType::Decimal(rhs_p, rhs_s)) => {
if lhs_p == rhs_p && lhs_s == rhs_s {
// Closure for the binary operation.
let op = move |a, b| {
let res: i128 = a + b;
let max = max_value(*lhs_p);
let (precision, _) = get_parameters(lhs.data_type(), rhs.data_type())?;

match res {
res if res.abs() > max => {
if res > 0 {
max
} else {
-max
}
}
_ => res,
}
};
let max = max_value(precision);
let op = move |a, b| {
let res: i128 = a + b;

binary(lhs, rhs, lhs.data_type().clone(), op)
if res.abs() > max {
if res > 0 {
max
} else {
Err(ArrowError::InvalidArgumentError(
"Arrays must have the same precision and scale".to_string(),
))
-max
}
} else {
res
}
_ => Err(ArrowError::InvalidArgumentError(
"Incorrect data type for the array".to_string(),
)),
}
};

binary(lhs, rhs, lhs.data_type().clone(), op)
}

/// Checked addition of two decimal primitive arrays with the same precision
Expand All @@ -171,33 +124,20 @@ pub fn checked_add(
lhs: &PrimitiveArray<i128>,
rhs: &PrimitiveArray<i128>,
) -> Result<PrimitiveArray<i128>> {
// Matching on both data types from both arrays. This match will be true
// only when precision and scale from both arrays are the same, otherwise
// it will return and ArrowError
match (lhs.data_type(), rhs.data_type()) {
(DataType::Decimal(lhs_p, lhs_s), DataType::Decimal(rhs_p, rhs_s)) => {
if lhs_p == rhs_p && lhs_s == rhs_s {
// Closure for the binary operation.
let op = move |a, b| {
let res: i128 = a + b;
let (precision, _) = get_parameters(lhs.data_type(), rhs.data_type())?;

match res {
res if res.abs() > max_value(*lhs_p) => None,
_ => Some(res),
}
};
let max = max_value(precision);
let op = move |a, b| {
let result: i128 = a + b;

binary_checked(lhs, rhs, lhs.data_type().clone(), op)
} else {
Err(ArrowError::InvalidArgumentError(
"Arrays must have the same precision and scale".to_string(),
))
}
if result.abs() > max {
None
} else {
Some(result)
}
_ => Err(ArrowError::InvalidArgumentError(
"Incorrect data type for the array".to_string(),
)),
}
};

binary_checked(lhs, rhs, lhs.data_type().clone(), op)
}

// Implementation of ArrayAdd trait for PrimitiveArrays
Expand Down Expand Up @@ -259,14 +199,16 @@ pub fn adaptive_add(
// looping through the iterator
let (mut res_p, res_s, diff) = adjusted_precision_scale(*lhs_p, *lhs_s, *rhs_p, *rhs_s);

let mut result = Vec::new();
for (l, r) in lhs.values().iter().zip(rhs.values().iter()) {
let shift = 10i128.pow(diff as u32);
let mut max = max_value(res_p);

let iter = lhs.values().iter().zip(rhs.values().iter()).map(|(l, r)| {
// Based on the array's scales one of the arguments in the sum has to be shifted
// to the left to match the final scale
let res = if lhs_s > rhs_s {
l + r * 10i128.pow(diff as u32)
l + r * shift
} else {
l * 10i128.pow(diff as u32) + r
l * shift + r
};

// The precision of the resulting array will change if one of the
Expand All @@ -277,15 +219,15 @@ pub fn adaptive_add(
// 00.0001 -> 6, 4
// -----------------
// 100.0000 -> 7, 4
if res.abs() > max_value(res_p) {
if res.abs() > max {
res_p = number_digits(res);
max = max_value(res_p);
}

result.push(res);
}
res
});
let values = Buffer::from_trusted_len_iter(iter);

let validity = combine_validities(lhs.validity(), rhs.validity());
let values = Buffer::from(result);

Ok(PrimitiveArray::<i128>::from_data(
DataType::Decimal(res_p, res_s),
Expand Down
Loading