paddyhoran commented on code in PR #264:
URL: https://github.com/apache/arrow-site/pull/264#discussion_r1011714420
##########
_posts/2022-10-30-multi-column-sorts-in-arrow-rust-part-1.md:
##########
@@ -0,0 +1,232 @@
+---
+layout: post
+title: "Fast and Memory Efficient Multi-Column Sorts in Apache Arrow Rust,
Part 1"
+date: "2022-10-30 00:00:00"
+author: "tustvold and alamb"
+categories: [arrow]
+---
+<!--
+{% comment %}
+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.
+{% endcomment %}
+-->
+
+## Introduction
+
+Sorting is one of the most fundamental operations in modern databases and
other analytic systems, underpinning important operators such as aggregates,
joins, window functions, merge, and more. By some estimates, more than half of
the execution time in data processing systems is spent sorting. Optimizing
sorts is therefore vital to improving query performance and overall system
efficiency.
+
+Sorting is also one of the most well studied topics in computer science. The
classic survey paper for databases is [Implementing Sorting in Database
Systems](https://dl.acm.org/doi/10.1145/1132960.1132964) by Goetz Graefe which
provides a thorough academic treatment and is still very applicable today.
However, it may not be obvious how to apply the wisdom and advanced techniques
described in that paper to modern systems. In addition, the excellent [DuckDB
blog on sorting](https://duckdb.org/2021/08/27/external-sorting.html)
highlights many sorting techniques, and mentions a comparable row format, but
it does not explain how to efficiently sort variable length strings or
dictionary encoded data.
+
+In this blog post we explain in detail the new [row
format](https://docs.rs/arrow/25.0.0/arrow/row/index.html) in the [Rust
implementation](https://github.com/apache/arrow-rs) of [Apache
Arrow](https://arrow.apache.org/), and how we used to make sorting more than
[3x](https://github.com/apache/arrow-rs/pull/2929) faster than an alternate
comparator based approach. The benefits are especially pronounced for strings,
dictionary encoded data, and sorts with large numbers of columns.
+
+
+## Multicolumn / Lexicographical Sort Problem
+
+Most languages have native, optimized operations to sort a single column
(array) of data, which are specialized based on the type of data being sorted.
The reason that sorting is typically more challenging in analytic systems is
that it must:
+
+1. Support multiple columns of data
+2. The column types are not knowable at compile time, and thus the compiler
can not typically generate optimized code.
+
+Multicolumn sorting is also referred to as lexicographical sorting in some
libraries.
+
+For example, given sales data for various customers and their state of
residence, a user might want to find the lowest 10 orders for each state.
+
+```text
+Customer | State | Orders
+—--------+-------+-------
+12345 | MA | 10.12
+532432 | MA | 8.44
+12345 | CA | 3.25
+56232 | WA | 6.00
+23442 | WA | 132.50
+7844 | CA | 9.33
+852353 | MA | 1.30
+```
+
+One way to do so is to order the data first by `State` and then by `Orders`:
+```text
+Customer | State | Orders
+—--------+-------+-------
+12345 | CA | 3.25
+7844 | CA | 9.33
+852353 | MA | 1.30
+532432 | MA | 8.44
+12345 | MA | 10.12
+56232 | WA | 6.00
+23442 | WA | 132.50
+```
+
+(Note: While there are specialized ways for computing this particular query
other than fully sorting the entire input such (e.g. "TopK"), they typically
need the same multi-column comparison operation described below. Thus while we
will use the simplified example in our post, it applies much more broadly)
Review Comment:
```suggestion
(Note: While there are specialized ways for computing this particular query
other than fully sorting the entire input (e.g. "TopK"), they typically need
the same multi-column comparison operation described below. Thus while we will
use the simplified example in our post, it applies much more broadly)
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
