utils.hpp

Go to the documentation of this file.

#pragma once
 
#include <cstdint>
#include <optional>
#include <string_view>
 
#include <sparrow/record_batch.hpp>
 
#include "sparrow_ipc/config/config.hpp"
 
namespace sparrow_ipc::utils
{
    // Aligns a value to the next multiple of 8, as required by the Arrow IPC format for message bodies
    constexpr size_t align_to_8(const size_t n)
    {
        return (n + 7) & -8;
    }

    SPARROW_IPC_API std::optional<std::string_view> get_substr_after_separator(std::string_view str, std::string_view sep);

    SPARROW_IPC_API std::optional<int32_t> parse_to_int32(std::string_view str);

    SPARROW_IPC_API std::optional<int32_t> get_substr_as_int32(std::string_view str, std::string_view sep);

    SPARROW_IPC_API std::optional<std::tuple<int32_t, int32_t, std::optional<int32_t>>> parse_decimal_format(std::string_view format_str);

    SPARROW_IPC_API std::vector<std::string_view> extract_words_after_colon(std::string_view str);

    template <std::ranges::input_range R>
        requires std::same_as<std::ranges::range_value_t<R>, sparrow::record_batch>
    bool check_record_batches_consistency(const R& record_batches)
    {
        if (record_batches.empty() || record_batches.size() == 1)
        {
            return true;
        }
        const sparrow::record_batch& first_rb = record_batches[0];
        const size_t first_rb_nb_columns = first_rb.nb_columns();
        for (const sparrow::record_batch& rb : record_batches)
        {
            const auto rb_nb_columns = rb.nb_columns();
            if (rb_nb_columns != first_rb_nb_columns)
            {
                return false;
            }
            for (size_t col_idx = 0; col_idx < rb.nb_columns(); ++col_idx)
            {
                const sparrow::array& arr = rb.get_column(col_idx);
                const sparrow::array& first_arr = first_rb.get_column(col_idx);
                const auto arr_data_type = arr.data_type();
                const auto first_arr_data_type = first_arr.data_type();
                if (arr_data_type != first_arr_data_type)
                {
                    return false;
                }
            }
        }
        return true;
    }
}