Function fillColumnWithRandomData

src/Storages/StorageGenerateRandom.cpp:83–400

ColumnPtr fillColumnWithRandomData(
    const DataTypePtr type,
    UInt64 limit,
    UInt64 max_array_length,
    UInt64 max_string_length,
    pcg64 & rng,
    ContextPtr context)
{
    TypeIndex idx = type->getTypeId();

    switch (idx)
    {
        case TypeIndex::String:
        {
            /// Mostly the same as the implementation of randomPrintableASCII function.

            auto column = ColumnString::create();
            ColumnString::Chars & data_to = column->getChars();
            ColumnString::Offsets & offsets_to = column->getOffsets();
            offsets_to.resize(limit);

            IColumn::Offset offset = 0;
            for (size_t row_num = 0; row_num < limit; ++row_num)
            {
                size_t length = rng() % (max_string_length + 1);    /// Slow

                IColumn::Offset next_offset = offset + length + 1;
                data_to.resize(next_offset);
                offsets_to[row_num] = next_offset;

                auto * data_to_ptr = data_to.data();    /// avoid assert on array indexing after end
                for (size_t pos = offset, end = offset + length; pos < end; pos += 4)    /// We have padding in column buffers that we can overwrite.
                {
                    UInt64 rand = rng();

                    UInt16 rand1 = rand;
                    UInt16 rand2 = rand >> 16;
                    UInt16 rand3 = rand >> 32;
                    UInt16 rand4 = rand >> 48;

                    /// Printable characters are from range [32; 126].
                    /// https://lemire.me/blog/2016/06/27/a-fast-alternative-to-the-modulo-reduction/

                    data_to_ptr[pos + 0] = 32 + ((rand1 * 95) >> 16);
                    data_to_ptr[pos + 1] = 32 + ((rand2 * 95) >> 16);
                    data_to_ptr[pos + 2] = 32 + ((rand3 * 95) >> 16);
                    data_to_ptr[pos + 3] = 32 + ((rand4 * 95) >> 16);

                    /// NOTE gcc failed to vectorize this code (aliasing of char?)
                    /// TODO Implement SIMD optimizations from Danila Kutenin.
                }

                data_to[offset + length] = 0;

                offset = next_offset;
            }

            return column;