void IAST::getTreeHashImpl(SipHash & hash_state) const
{
auto id = getID();
hash_state.update(id.data(), id.size());
size_t num_children = children.size();
hash_state.update(reinterpret_cast<const char *>(&num_children), sizeof(num_children));
for (const auto & child : children)
child->getTreeHashImpl(hash_state);
}
static Compiler::HashedKey getHash(const std::string & key)
{
SipHash hash;
auto revision = ClickHouseRevision::get();
hash.update(reinterpret_cast<const char *>(&revision), sizeof(revision));
hash.update(key.data(), key.size());
Compiler::HashedKey res;
hash.get128(res.first, res.second);
return res;
}
void IAST::updateTreeHash(SipHash & hash_state) const
{
updateTreeHashImpl(hash_state);
hash_state.update(children.size());
for (const auto & child : children)
child->updateTreeHash(hash_state);
}
void ColumnArray::updateHashWithValue(size_t n, SipHash & hash) const
{
size_t array_size = sizeAt(n);
size_t offset = offsetAt(n);
hash.update(array_size);
for (size_t i = 0; i < array_size; ++i)
getData().updateHashWithValue(offset + i, hash);
}
void IAST::updateTreeHashImpl(SipHash & hash_state) const
{
auto id = getID();
hash_state.update(id.data(), id.size());
}
void ColumnFixedString::updateHashWithValue(size_t index, SipHash & hash) const
{
hash.update(reinterpret_cast<const char *>(&chars[n * index]), n);
}
void ColumnVector<T>::updateHashWithValue(size_t n, SipHash & hash) const
{
hash.update(data[n]);
}
int main(int argc, char ** argv)
{
using Strings = std::vector<std::string>;
using Hashes = std::vector<char>;
Strings strings;
size_t rows = 0;
size_t bytes = 0;
{
Stopwatch watch;
DB::ReadBufferFromFileDescriptor in(STDIN_FILENO);
while (!in.eof())
{
strings.push_back(std::string());
DB::readEscapedString(strings.back(), in);
DB::assertChar('\n', in);
bytes += strings.back().size() + 1;
}
watch.stop();
rows = strings.size();
std::cerr << std::fixed << std::setprecision(2)
<< "Read " << rows << " rows, " << bytes / 1000000.0 << " MB"
<< ", elapsed: " << watch.elapsedSeconds()
<< " (" << rows / watch.elapsedSeconds() << " rows/sec., " << bytes / 1000000.0 / watch.elapsedSeconds() << " MB/sec.)"
<< std::endl;
}
Hashes hashes(16 * rows);
{
Stopwatch watch;
for (size_t i = 0; i < rows; ++i)
{
*reinterpret_cast<UInt64*>(&hashes[i * 16]) = CityHash64(strings[i].data(), strings[i].size());
}
watch.stop();
UInt64 check = CityHash64(&hashes[0], hashes.size());
std::cerr << std::fixed << std::setprecision(2)
<< "CityHash64 (check = " << check << ")"
<< ", elapsed: " << watch.elapsedSeconds()
<< " (" << rows / watch.elapsedSeconds() << " rows/sec., " << bytes / 1000000.0 / watch.elapsedSeconds() << " MB/sec.)"
<< std::endl;
}
/* {
Stopwatch watch;
std::vector<char> seed(16);
for (size_t i = 0; i < rows; ++i)
{
sipHash(
reinterpret_cast<unsigned char *>(&hashes[i * 16]),
reinterpret_cast<const unsigned char *>(strings[i].data()),
strings[i].size(),
reinterpret_cast<const unsigned char *>(&seed[0]));
}
watch.stop();
UInt64 check = CityHash64(&hashes[0], hashes.size());
std::cerr << std::fixed << std::setprecision(2)
<< "SipHash (check = " << check << ")"
<< ", elapsed: " << watch.elapsedSeconds()
<< " (" << rows / watch.elapsedSeconds() << " rows/sec., " << bytes / 1000000.0 / watch.elapsedSeconds() << " MB/sec.)"
<< std::endl;
}*/
{
Stopwatch watch;
for (size_t i = 0; i < rows; ++i)
{
SipHash hash;
hash.update(strings[i].data(), strings[i].size());
hash.get128(&hashes[i * 16]);
}
watch.stop();
UInt64 check = CityHash64(&hashes[0], hashes.size());
std::cerr << std::fixed << std::setprecision(2)
<< "SipHash, stream (check = " << check << ")"
<< ", elapsed: " << watch.elapsedSeconds()
<< " (" << rows / watch.elapsedSeconds() << " rows/sec., " << bytes / 1000000.0 / watch.elapsedSeconds() << " MB/sec.)"
<< std::endl;
}
{
Stopwatch watch;
for (size_t i = 0; i < rows; ++i)
{
MD5_CTX state;
MD5_Init(&state);
MD5_Update(&state, reinterpret_cast<const unsigned char *>(strings[i].data()), strings[i].size());
MD5_Final(reinterpret_cast<unsigned char *>(&hashes[i * 16]), &state);
}
watch.stop();
UInt64 check = CityHash64(&hashes[0], hashes.size());
std::cerr << std::fixed << std::setprecision(2)
<< "MD5 (check = " << check << ")"
<< ", elapsed: " << watch.elapsedSeconds()
<< " (" << rows / watch.elapsedSeconds() << " rows/sec., " << bytes / 1000000.0 / watch.elapsedSeconds() << " MB/sec.)"
<< std::endl;
}
return 0;
}
