Rich-T-kid commented on code in PR #21765: URL: https://github.com/apache/datafusion/pull/21765#discussion_r3373379676
########## datafusion/physical-plan/src/aggregates/group_values/single_group_by/dictionary.rs: ########## @@ -0,0 +1,878 @@ +// 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 crate::aggregates::group_values::GroupValues; +use crate::hash_utils::RandomState; +use arrow::array::{ + Array, ArrayRef, AsArray, DictionaryArray, LargeStringArray, LargeStringBuilder, + ListArray, ListBuilder, PrimitiveArray, PrimitiveBuilder, StringArray, StringBuilder, + StringViewArray, StringViewBuilder, +}; +use arrow::datatypes::{ArrowDictionaryKeyType, ArrowNativeType, DataType}; +use datafusion_common::DataFusionError::{Internal, NotImplemented}; +use datafusion_common::Result; +use datafusion_common::hash_utils::create_hashes; +use datafusion_expr::EmitTo; +use hashbrown::HashTable; +use hashbrown::hash_table::Entry as HashTableEntry; +use std::borrow::Cow; +use std::marker::PhantomData; +use std::sync::Arc; + +/// Heuristic for sizing the values buffer of string builders during emit: +/// dictionary-encoded values are short by design (categorical strings, short +/// identifiers), so 16 B/item avoids the realloc-doubling chain in the common +/// case while keeping over-allocation cheap when values are smaller. +const AVG_BYTES_PER_DICT_VALUE: usize = 16; +const INITIAL_PRE_ALLOCATION: usize = 8 * 1024; // avoid re-allocation`s for small-medium groups Review Comment: ideally the 'cheap signal' would be the first dictionary arrays values, assuming that the values are unique. the tables capacity is pre-allocated at creation time not when `intern()` is first called so I think it makes sense to just remove the pre-allocation all together and let it grow. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
