alamb commented on code in PR #12047: URL: https://github.com/apache/datafusion/pull/12047#discussion_r1720968755
########## datafusion/physical-plan/src/coalesce/mod.rs: ########## @@ -0,0 +1,588 @@ +// 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. + +use arrow::compute::concat_batches; +use arrow_array::builder::StringViewBuilder; +use arrow_array::cast::AsArray; +use arrow_array::{Array, ArrayRef, RecordBatch}; +use arrow_schema::SchemaRef; +use std::sync::Arc; + +/// Concatenate multiple [`RecordBatch`]es +/// +/// `BatchCoalescer` concatenates multiple small [`RecordBatch`]es, produced by +/// operations such as `FilterExec` and `RepartitionExec`, into larger ones for +/// more efficient processing by subsequent operations. +/// +/// # Background +/// +/// Generally speaking, larger [`RecordBatch`]es are more efficient to process +/// than smaller record batches (until the CPU cache is exceeded) because there +/// is fixed processing overhead per batch. DataFusion tries to operate on +/// batches of `target_batch_size` rows to amortize this overhead +/// +/// ```text +/// ┌────────────────────┐ +/// │ RecordBatch │ +/// │ num_rows = 23 │ +/// └────────────────────┘ ┌────────────────────┐ +/// │ │ +/// ┌────────────────────┐ Coalesce │ │ +/// │ │ Batches │ │ +/// │ RecordBatch │ │ │ +/// │ num_rows = 50 │ ─ ─ ─ ─ ─ ─ ▶ │ │ +/// │ │ │ RecordBatch │ +/// │ │ │ num_rows = 106 │ +/// └────────────────────┘ │ │ +/// │ │ +/// ┌────────────────────┐ │ │ +/// │ │ │ │ +/// │ RecordBatch │ │ │ +/// │ num_rows = 33 │ └────────────────────┘ +/// │ │ +/// └────────────────────┘ +/// ``` +/// +/// # Notes: +/// +/// 1. Output rows are produced in the same order as the input rows +/// +/// 2. The output is a sequence of batches, with all but the last being at least +/// `target_batch_size` rows. +/// +/// 3. Eventually this may also be able to handle other optimizations such as a +/// combined filter/coalesce operation. +/// +#[derive(Debug)] +pub struct BatchCoalescer { Review Comment: The only changes in this module are to make the struct's and some methods `pub` and update the documentation slightly ########## datafusion/physical-plan/src/coalesce_batches.rs: ########## @@ -346,7 +316,7 @@ impl CoalesceBatchesStream { } CoalesceBatchesStreamState::Exhausted => { // Handle the end of the input stream. - return if self.coalescer.buffer.is_empty() { + return if self.coalescer.is_empty() { Review Comment: Putting the BatchCoalescer into its own module means this code must use an accessor rather than directly look at the coalescer's state, which I think improves the modularity (a tiny bit) -- 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: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
