jorgecarleitao commented on a change in pull request #8709: URL: https://github.com/apache/arrow/pull/8709#discussion_r527880432
########## File path: rust/datafusion/src/physical_plan/hash_join.rs ########## @@ -0,0 +1,507 @@ +// 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 join plan for executing partitions in parallel and then joining the results +//! into a set of partitions. + +use std::sync::Arc; +use std::{ + any::Any, + collections::{HashMap, HashSet}, +}; + +use async_trait::async_trait; +use futures::{Stream, StreamExt, TryStreamExt}; + +use arrow::array::{make_array, Array, MutableArrayData}; +use arrow::datatypes::{Schema, SchemaRef}; +use arrow::error::Result as ArrowResult; +use arrow::record_batch::RecordBatch; + +use super::{expressions::col, hash_aggregate::create_key}; +use super::{ + hash_utils::{build_join_schema, check_join_is_valid, JoinType}, + merge::MergeExec, +}; +use crate::error::{DataFusionError, Result}; + +use super::{ + group_scalar::GroupByScalar, ExecutionPlan, Partitioning, RecordBatchStream, + SendableRecordBatchStream, +}; + +// An index of (batch, row) uniquely identifying a row in a part. +type Index = (usize, usize); +// A pair (left index, right index) +// Note that while this is currently equal to `Index`, the `JoinIndex` is semantically different +// as a left join may issue None indices, in which case +type JoinIndex = Option<(usize, usize)>; +// Maps ["on" value] -> [list of indices with this key's value] +// E.g. [1, 2] -> [(0, 3), (1, 6), (0, 8)] indicates that (column1, column2) = [1, 2] is true +// for rows 3 and 8 from batch 0 and row 6 from batch 1. +type JoinHashMap = HashMap<Vec<GroupByScalar>, Vec<Index>>; +type JoinLeftData = (JoinHashMap, Vec<RecordBatch>); + +/// join execution plan executes partitions in parallel and combines them into a set of +/// partitions. +#[derive(Debug)] +pub struct HashJoinExec { + /// left side Review comment: @alamb thanks a lot for the review. Really great knowledge and notation about this. I changed the wording to match that. My largest concern (which I think you also had) was that this PR was changing the `filter` code. I removed all those changes to have this PR focused on the `join`, and not force us to change the `filter` because we wanted this. I agree with your comment on the other PR that we may want to keep the `filter` code as is. The change of making the `MutableDataArray` accepting multiple arrays makes the filter even slower, which IMO further strengths that idea. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
