arrow/csv/column_builder.h

0001 // Licensed to the Apache Software Foundation (ASF) under one
0002 // or more contributor license agreements.  See the NOTICE file
0003 // distributed with this work for additional information
0004 // regarding copyright ownership.  The ASF licenses this file
0005 // to you under the Apache License, Version 2.0 (the
0006 // "License"); you may not use this file except in compliance
0007 // with the License.  You may obtain a copy of the License at
0008 //
0009 //   http://www.apache.org/licenses/LICENSE-2.0
0010 //
0011 // Unless required by applicable law or agreed to in writing,
0012 // software distributed under the License is distributed on an
0013 // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
0014 // KIND, either express or implied.  See the License for the
0015 // specific language governing permissions and limitations
0016 // under the License.
0017
0018 #pragma once
0019
0020 #include <cstdint>
0021 #include <memory>
0022 #include <utility>
0023
0024 #include "arrow/result.h"
0025 #include "arrow/type_fwd.h"
0026 #include "arrow/util/type_fwd.h"
0027 #include "arrow/util/visibility.h"
0028
0029 namespace arrow {
0030 namespace csv {
0031
0032 class BlockParser;
0033 struct ConvertOptions;
0034
0035 class ARROW_EXPORT ColumnBuilder {
0036  public:
0037   virtual ~ColumnBuilder() = default;
0038
0039   /// Spawn a task that will try to convert and append the given CSV block.
0040   /// All calls to Append() should happen on the same thread, otherwise
0041   /// call Insert() instead.
0042   virtual void Append(const std::shared_ptr<BlockParser>& parser) = 0;
0043
0044   /// Spawn a task that will try to convert and insert the given CSV block
0045   virtual void Insert(int64_t block_index,
0046                       const std::shared_ptr<BlockParser>& parser) = 0;
0047
0048   /// Return the final chunked array.  The TaskGroup _must_ have finished!
0049   virtual Result<std::shared_ptr<ChunkedArray>> Finish() = 0;
0050
0051   std::shared_ptr<arrow::internal::TaskGroup> task_group() { return task_group_; }
0052
0053   /// Construct a strictly-typed ColumnBuilder.
0054   static Result<std::shared_ptr<ColumnBuilder>> Make(
0055       MemoryPool* pool, const std::shared_ptr<DataType>& type, int32_t col_index,
0056       const ConvertOptions& options,
0057       const std::shared_ptr<arrow::internal::TaskGroup>& task_group);
0058
0059   /// Construct a type-inferring ColumnBuilder.
0060   static Result<std::shared_ptr<ColumnBuilder>> Make(
0061       MemoryPool* pool, int32_t col_index, const ConvertOptions& options,
0062       const std::shared_ptr<arrow::internal::TaskGroup>& task_group);
0063
0064   /// Construct a ColumnBuilder for a column of nulls
0065   /// (i.e. not present in the CSV file).
0066   static Result<std::shared_ptr<ColumnBuilder>> MakeNull(
0067       MemoryPool* pool, const std::shared_ptr<DataType>& type,
0068       const std::shared_ptr<arrow::internal::TaskGroup>& task_group);
0069
0070  protected:
0071   explicit ColumnBuilder(std::shared_ptr<arrow::internal::TaskGroup> task_group)
0072       : task_group_(std::move(task_group)) {}
0073
0074   std::shared_ptr<arrow::internal::TaskGroup> task_group_;
0075 };
0076
0077 }  // namespace csv
0078 }  // namespace arrow