Rich-T-kid commented on code in PR #23187: URL: https://github.com/apache/datafusion/pull/23187#discussion_r3506802100
########## datafusion/physical-plan/src/aggregates/group_values/multi_group_by/dictionary.rs: ########## @@ -0,0 +1,490 @@ +// 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::multi_group_by::GroupColumn; + +use arrow::array::{ + Array, ArrayRef, AsArray, BooleanBufferBuilder, DictionaryArray, PrimitiveArray, +}; +use arrow::datatypes::{ArrowDictionaryKeyType, ArrowNativeType, Field}; +use datafusion_common::Result; +use std::marker::PhantomData; +use std::sync::Arc; + +pub struct DictionaryGroupValuesColumn<K: ArrowDictionaryKeyType + Send + Sync> { + inner: Box<dyn GroupColumn>, + null_array: ArrayRef, + _phantom: PhantomData<K>, +} + +impl<K: ArrowDictionaryKeyType + Send + Sync> DictionaryGroupValuesColumn<K> { + pub fn new(inner: Box<dyn GroupColumn>, field: &Field) -> Self { + let null_array = arrow::array::new_null_array(field.data_type(), 1); + Self { + inner, + null_array, + _phantom: PhantomData, + } + } + + #[inline] + fn into_dict(values: ArrayRef) -> ArrayRef { + // at some point in the future we may want to support bumping the key types + // https://github.com/apache/datafusion/issues/23127 + let keys: PrimitiveArray<K> = (0..values.len()) + .map(|i| { + if values.is_null(i) { + None + } else { + Some(K::Native::usize_as(i)) + } + }) + .collect(); + Arc::new(DictionaryArray::<K>::new(keys, values)) + } +} + +impl<K: ArrowDictionaryKeyType + Send + Sync> GroupColumn + for DictionaryGroupValuesColumn<K> +{ + fn equal_to(&self, lhs_row: usize, array: &ArrayRef, rhs_row: usize) -> bool { + let dict = array.as_dictionary::<K>(); + match dict.key(rhs_row) { + None => self.inner.equal_to(lhs_row, &self.null_array, 0), + Some(key) => self.inner.equal_to(lhs_row, dict.values(), key), + } + } + + fn append_val(&mut self, array: &ArrayRef, row: usize) -> Result<()> { + let dict = array.as_dictionary::<K>(); + match dict.key(row) { + None => self.inner.append_val(&self.null_array, 0), + Some(key) => self.inner.append_val(dict.values(), key), + } + } + + fn vectorized_equal_to( + &self, + lhs_rows: &[usize], + array: &ArrayRef, + rhs_rows: &[usize], + equal_to_results: &mut BooleanBufferBuilder, + ) { + let dict = array.as_dictionary::<K>(); + let keys = dict.keys(); + let values = dict.values(); + + if keys.null_count() == 0 { + let key_indices: Vec<usize> = + rhs_rows.iter().map(|&r| keys.value(r).as_usize()).collect(); + self.inner.vectorized_equal_to( + lhs_rows, + values, + &key_indices, + equal_to_results, + ); + } else { + for (i, (lhs_row, rhs_row)) in lhs_rows.iter().zip(rhs_rows).enumerate() { + if equal_to_results.get_bit(i) { + let result = match dict.key(*rhs_row) { + None => self.inner.equal_to(*lhs_row, &self.null_array, 0), + Some(key) => self.inner.equal_to(*lhs_row, values, key), + }; + equal_to_results.set_bit(i, result); + } + } + } + } + + fn vectorized_append(&mut self, array: &ArrayRef, rows: &[usize]) -> Result<()> { + let dict = array.as_dictionary::<K>(); + let keys = dict.keys(); + + if keys.null_count() == 0 { + let key_indices: Vec<usize> = + rows.iter().map(|&r| keys.value(r).as_usize()).collect(); + self.inner.vectorized_append(dict.values(), &key_indices) + } else { + let values = dict.values(); + for &row in rows { + match dict.key(row) { + None => self.inner.append_val(&self.null_array, 0)?, + Some(k) => self.inner.append_val(values, k)?, + } + } + Ok(()) Review Comment: `take` is slower than doing per-row work. Another possible solution is to build append a null array to the end and re-map the keys that are null -- 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]
