sunchao commented on code in PR #24526: URL: https://github.com/apache/datafusion/pull/24526#discussion_r3866800177
########## datafusion/pruning/src/string_in_list.rs: ########## @@ -0,0 +1,225 @@ +// 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 std::fmt::{self, Display, Formatter}; +use std::hash::{Hash, Hasher}; +use std::sync::Arc; + +use arrow::array::{Array, AsArray, BooleanArray}; +use arrow::compute::cast; +use arrow::datatypes::{DataType, Schema}; +use arrow::record_batch::RecordBatch; +use datafusion_common::{Result, assert_eq_or_internal_err}; +use datafusion_physical_expr::{PhysicalExpr, PhysicalExprRef}; +use datafusion_physical_plan::ColumnarValue; + +/// Tests whether a sorted string domain intersects an inclusive statistics interval. +/// This expression is used only for pruning; the original IN remains the row filter. +#[derive(Debug, Eq)] +pub(crate) struct StringInListPruningExpr { + min: PhysicalExprRef, + max: PhysicalExprRef, + values: Arc<[String]>, Review Comment: Not yet: the current benchmark compares compact pruning with the per-value expression tree, not alternative domain representations. `Arc<[String]>` makes sorting/deduplication straightforward and lets `with_new_children` share the prepared domain. `StringArray`/`LargeStringArray` would pack the bytes and offsets contiguously and is worth comparing for construction cost, allocations, and lookup locality. I don't have evidence that the current representation is optimal; I'd keep that measurement and any resulting change as a follow-up. ########## datafusion/pruning/src/string_in_list.rs: ########## @@ -0,0 +1,225 @@ +// 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 std::fmt::{self, Display, Formatter}; +use std::hash::{Hash, Hasher}; +use std::sync::Arc; + +use arrow::array::{Array, AsArray, BooleanArray}; +use arrow::compute::cast; +use arrow::datatypes::{DataType, Schema}; +use arrow::record_batch::RecordBatch; +use datafusion_common::{Result, assert_eq_or_internal_err}; +use datafusion_physical_expr::{PhysicalExpr, PhysicalExprRef}; +use datafusion_physical_plan::ColumnarValue; + +/// Tests whether a sorted string domain intersects an inclusive statistics interval. +/// This expression is used only for pruning; the original IN remains the row filter. +#[derive(Debug, Eq)] +pub(crate) struct StringInListPruningExpr { + min: PhysicalExprRef, + max: PhysicalExprRef, + values: Arc<[String]>, +} + +impl StringInListPruningExpr { + pub(crate) fn new( + min: PhysicalExprRef, + max: PhysicalExprRef, + mut values: Vec<String>, + ) -> Self { + values.sort_unstable(); + values.dedup(); + Self { + min, + max, + values: values.into(), + } + } +} + +impl PartialEq for StringInListPruningExpr { + fn eq(&self, other: &Self) -> bool { + self.min.eq(&other.min) && self.max.eq(&other.max) && self.values == other.values + } +} + +impl Hash for StringInListPruningExpr { + fn hash<H: Hasher>(&self, state: &mut H) { + self.min.hash(state); + self.max.hash(state); + self.values.hash(state); + } +} Review Comment: Agreed. Hashing traverses the domain, and equality can do so too. The existing benchmark measures complete predicate construction and pruning evaluation; it does not isolate repeated hashing/equality during expression rewrites. I'll leave this unchanged here as suggested. A follow-up should measure large and long-string domains; a cached digest could avoid repeated hashing, while equality must still compare values when digests match so collisions cannot equate different predicates. -- 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]
