yordan-pavlov commented on a change in pull request #384: URL: https://github.com/apache/arrow-rs/pull/384#discussion_r646118671
########## File path: parquet/src/arrow/arrow_array_reader.rs ########## @@ -0,0 +1,1394 @@ +// 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::{any::Any, collections::VecDeque, marker::PhantomData}; +use std::{rc::Rc, cell::RefCell}; +use arrow::{array::{ArrayRef, Int16Array}, buffer::MutableBuffer, datatypes::{DataType as ArrowType, ToByteSlice}}; +use crate::{column::page::{Page, PageIterator}, memory::ByteBufferPtr, schema::types::{ColumnDescPtr, ColumnDescriptor}}; +use crate::arrow::schema::parquet_to_arrow_field; +use crate::errors::{ParquetError, Result}; +use crate::basic::Encoding; +use super::array_reader::ArrayReader; + +struct UnzipIter<Source, Target, State> +{ + shared_state: Rc<RefCell<State>>, + select_item_buffer: fn(&mut State) -> &mut VecDeque<Target>, + consume_source_item: fn(source_item: Source, state: &mut State) -> Target, +} + +impl<Source, Target, State> UnzipIter<Source, Target, State> +{ + fn new( + shared_state: Rc<RefCell<State>>, + item_buffer_selector: fn(&mut State) -> &mut VecDeque<Target>, + source_item_consumer: fn(source_item: Source, state: &mut State) -> Target + ) -> Self { + Self { + shared_state, + select_item_buffer: item_buffer_selector, + consume_source_item: source_item_consumer, + } + } +} + +trait UnzipIterState<T> { + type SourceIter: Iterator<Item = T>; + fn source_iter(&mut self) -> &mut Self::SourceIter; +} + +impl<Source, Target, State: UnzipIterState<Source>> Iterator for UnzipIter<Source, Target, State> { + type Item = Target; + + fn next(&mut self) -> Option<Self::Item> { + let mut inner = self.shared_state.borrow_mut(); + // try to get one from the stored data + (self.select_item_buffer)(&mut *inner).pop_front().or_else(|| + // nothing stored, we need a new element. + inner.source_iter().next().map(|s| { + (self.consume_source_item)(s, &mut inner) + })) + } +} + +struct PageBufferUnzipIterState<V, L, It> { + iter: It, + value_iter_buffer: VecDeque<V>, + def_level_iter_buffer: VecDeque<L>, + rep_level_iter_buffer: VecDeque<L>, +} + +impl<V, L, It: Iterator<Item = (V, L, L)>> UnzipIterState<(V, L, L)> for PageBufferUnzipIterState<V, L, It> { + type SourceIter = It; + + #[inline] + fn source_iter(&mut self) -> &mut Self::SourceIter { + &mut self.iter + } +} + +fn unzip_iter<V, L, It: Iterator<Item = (V, L, L)>>(it: It) -> ( + UnzipIter<(V, L, L), V, PageBufferUnzipIterState<V, L, It>>, + UnzipIter<(V, L, L), L, PageBufferUnzipIterState<V, L, It>>, + UnzipIter<(V, L, L), L, PageBufferUnzipIterState<V, L, It>>, +) { + let shared_data = Rc::new(RefCell::new(PageBufferUnzipIterState { + iter: it, + value_iter_buffer: VecDeque::new(), + def_level_iter_buffer: VecDeque::new(), + rep_level_iter_buffer: VecDeque::new(), + })); + + let value_iter = UnzipIter::new( + shared_data.clone(), + |state| &mut state.value_iter_buffer, + |(v, d, r), state| { + state.def_level_iter_buffer.push_back(d); + state.rep_level_iter_buffer.push_back(r); + v + }, + ); + + let def_level_iter = UnzipIter::new( + shared_data.clone(), + |state| &mut state.def_level_iter_buffer, + |(v, d, r), state| { + state.value_iter_buffer.push_back(v); + state.rep_level_iter_buffer.push_back(r); + d + }, + ); + + let rep_level_iter = UnzipIter::new( + shared_data, + |state| &mut state.rep_level_iter_buffer, + |(v, d, r), state| { + state.value_iter_buffer.push_back(v); + state.def_level_iter_buffer.push_back(d); + r + }, + ); + + (value_iter, def_level_iter, rep_level_iter) +} + +pub trait ArrayConverter { + fn convert_value_bytes(&self, value_decoder: &mut impl ValueDecoder, num_values: usize) -> Result<arrow::array::ArrayData>; +} + +pub struct ArrowArrayReader<'a, C: ArrayConverter + 'a> { + column_desc: ColumnDescPtr, + data_type: ArrowType, + def_level_decoder: Box<dyn ValueDecoder + 'a>, + rep_level_decoder: Box<dyn ValueDecoder + 'a>, + value_decoder: Box<dyn ValueDecoder + 'a>, + last_def_levels: Option<Int16Array>, + last_rep_levels: Option<Int16Array>, + array_converter: C, +} + +pub(crate) struct ColumnChunkContext { + dictionary_values: Option<Vec<ByteBufferPtr>>, +} + +impl ColumnChunkContext { + fn new() -> Self { + Self { + dictionary_values: None, + } + } + + fn set_dictionary(&mut self, dictionary_values: Vec<ByteBufferPtr>) { + self.dictionary_values = Some(dictionary_values); + } +} + +impl<'a, C: ArrayConverter + 'a> ArrowArrayReader<'a, C> { + pub fn try_new<P: PageIterator + 'a>(column_chunk_iterator: P, column_desc: ColumnDescPtr, array_converter: C, arrow_type: Option<ArrowType>) -> Result<Self> { + let data_type = match arrow_type { + Some(t) => t, + None => parquet_to_arrow_field(column_desc.as_ref())? + .data_type() + .clone(), + }; + // println!("ArrowArrayReader::try_new, column: {}, data_type: {}", column_desc.path(), data_type); Review comment: done - comment removed -- 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]
