ColumnString improve performance(26%) by avoiding vector reallocate

[1261385937]

By estimating the total element count, code like this:

constexpr size_t total = 100000;
for (int i = 0; i < 10; i++) {
    clickhouse::Block block;
    auto a_ptr = std::make_shared<clickhouse::ColumnInt32>();
    auto b_ptr = std::make_shared<clickhouse::ColumnUInt64>();
    auto c_ptr = std::make_shared<clickhouse::ColumnString>(total);
    auto d_ptr = std::make_shared<clickhouse::ColumnArray>(std::make_shared<clickhouse::ColumnString>(total * 3));
    auto e_ptr = std::make_shared<clickhouse::ColumnArray>(
                         std::make_shared<clickhouse::ColumnArray>(
                         std::make_shared<clickhouse::ColumnString>(total * 3 * 4)));
...

Before estimating, total time(18720), network time(7353), other time(18720-7353=11367), top hot:

After estimating, total time(15728), network time(7317) same as above(7353), other time(15728-7317=8411), top hot:

improve performance: (11367-8411)/11367 = 26%