You've already forked mariadb-columnstore-engine
mirror of
https://github.com/mariadb-corporation/mariadb-columnstore-engine.git
synced 2025-07-30 19:23:07 +03:00
Update on review comments
This commit is contained in:
committed by
Leonid Fedorov
parent
865cca11c9
commit
0747099456
@ -107,7 +107,6 @@ void TableInfo::sleepMS(long ms)
|
||||
abs_ts.tv_sec = rm_ts.tv_sec;
|
||||
abs_ts.tv_nsec = rm_ts.tv_nsec;
|
||||
} while (nanosleep(&abs_ts, &rm_ts) < 0);
|
||||
|
||||
}
|
||||
|
||||
//------------------------------------------------------------------------------
|
||||
@ -156,7 +155,7 @@ TableInfo::TableInfo(Log* logger, const BRM::TxnID txnID, const string& processN
|
||||
, fRejectErrCnt(0)
|
||||
, fExtentStrAlloc(tableOID, logger)
|
||||
, fOamCachePtr(oam::OamCache::makeOamCache())
|
||||
, fParquetReader(NULL)
|
||||
, fParquetReader(nullptr)
|
||||
, fReader(nullptr)
|
||||
{
|
||||
fBuffers.clear();
|
||||
@ -271,8 +270,8 @@ int TableInfo::readTableData()
|
||||
{
|
||||
RID validTotalRows = 0;
|
||||
RID totalRowsPerInputFile = 0;
|
||||
int64_t totalRowsParquet = 0; // totalRowsParquet to be used in later function
|
||||
// needs int64_t type
|
||||
int64_t totalRowsParquet = 0; // totalRowsParquet to be used in later function
|
||||
// needs int64_t type
|
||||
int filesTBProcessed = fLoadFileList.size();
|
||||
int fileCounter = 0;
|
||||
unsigned long long qtSentAt = 0;
|
||||
@ -308,7 +307,6 @@ int TableInfo::readTableData()
|
||||
}
|
||||
fileCounter++;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
timeval readStart;
|
||||
@ -562,7 +560,8 @@ int TableInfo::readTableData()
|
||||
fCurrentReadBuffer = (fCurrentReadBuffer + 1) % fReadBufCount;
|
||||
|
||||
// bufferCount++;
|
||||
if ((fHandle && feof(fHandle)) || (fReadFromS3 && (fS3ReadLength == fS3ParseLength)) || (totalRowsPerInputFile == (RID)totalRowsParquet))
|
||||
if ((fHandle && feof(fHandle)) || (fReadFromS3 && (fS3ReadLength == fS3ParseLength)) ||
|
||||
(totalRowsPerInputFile == (RID)totalRowsParquet))
|
||||
{
|
||||
timeval readFinished;
|
||||
gettimeofday(&readFinished, NULL);
|
||||
@ -1147,8 +1146,7 @@ int TableInfo::getColumnForParse(const int& id, const int& bufferId, bool report
|
||||
|
||||
if (report)
|
||||
{
|
||||
oss << " ----- " << pthread_self() << ":fBuffers[" << bufferId <<
|
||||
"]: (colLocker,status,lasttime)- ";
|
||||
oss << " ----- " << pthread_self() << ":fBuffers[" << bufferId << "]: (colLocker,status,lasttime)- ";
|
||||
}
|
||||
|
||||
// @bug2099-
|
||||
@ -1232,8 +1230,8 @@ bool TableInfo::bufferReadyForParse(const int& bufferId, bool report) const
|
||||
ostringstream oss;
|
||||
string bufStatusStr;
|
||||
ColumnInfo::convertStatusToString(stat, bufStatusStr);
|
||||
oss << " --- " << pthread_self() <<
|
||||
":fBuffers[" << bufferId << "]=" << bufStatusStr << " (" << stat << ")" << std::endl;
|
||||
oss << " --- " << pthread_self() << ":fBuffers[" << bufferId << "]=" << bufStatusStr << " (" << stat
|
||||
<< ")" << std::endl;
|
||||
cout << oss.str();
|
||||
}
|
||||
|
||||
@ -1249,7 +1247,6 @@ bool TableInfo::bufferReadyForParse(const int& bufferId, bool report) const
|
||||
int TableInfo::initializeBuffers(int noOfBuffers, const JobFieldRefList& jobFieldRefList,
|
||||
unsigned int fixedBinaryRecLen)
|
||||
{
|
||||
|
||||
fReadBufCount = noOfBuffers;
|
||||
|
||||
// initialize and populate the buffer vector.
|
||||
@ -1292,8 +1289,7 @@ void TableInfo::addColumn(ColumnInfo* info)
|
||||
fExtentStrAlloc.addColumn(info->column.mapOid, info->column.width, info->column.dataType);
|
||||
}
|
||||
|
||||
|
||||
int TableInfo::openTableFileParquet(int64_t &totalRowsParquet)
|
||||
int TableInfo::openTableFileParquet(int64_t& totalRowsParquet)
|
||||
{
|
||||
if (fParquetReader != NULL)
|
||||
return NO_ERROR;
|
||||
@ -1323,14 +1319,13 @@ int TableInfo::openTableFileParquet(int64_t &totalRowsParquet)
|
||||
return ERR_FILE_OPEN;
|
||||
}
|
||||
// initialize fBuffers batch source
|
||||
for (int i = 0; i < fReadBufCount; ++i)
|
||||
for (auto& buffer : fBuffers)
|
||||
{
|
||||
fBuffers[i].setParquetReader(fParquetReader);
|
||||
buffer.setParquetReader(fParquetReader);
|
||||
}
|
||||
return NO_ERROR;
|
||||
|
||||
}
|
||||
|
||||
|
||||
//------------------------------------------------------------------------------
|
||||
// Open the file corresponding to fFileName so that we can import it's contents.
|
||||
// A buffer is also allocated and passed to setvbuf().
|
||||
@ -1424,7 +1419,7 @@ void TableInfo::closeTableFile()
|
||||
fclose(fHandle);
|
||||
delete[] fFileBuffer;
|
||||
}
|
||||
|
||||
|
||||
fHandle = 0;
|
||||
}
|
||||
else if (ms3)
|
||||
|
@ -35,8 +35,6 @@
|
||||
#include <iostream>
|
||||
using namespace std;
|
||||
|
||||
|
||||
|
||||
#include "bytestream.h"
|
||||
#include "brmtypes.h"
|
||||
#include "extentmap.h" // for DICT_COL_WIDTH
|
||||
@ -54,10 +52,13 @@ using namespace BRM;
|
||||
|
||||
using namespace idbdatafile;
|
||||
#include "checks.h"
|
||||
#include "utils_utf8.h" // for utf8_truncate_point()
|
||||
#include "utils_utf8.h" // for utf8_truncate_point()
|
||||
|
||||
namespace
|
||||
{
|
||||
using BinaryArraySharedPtr = std::shared_ptr<arrow::BinaryArray>;
|
||||
using FixedSizeBinaryArraySharedPtr = std::shared_ptr<arrow::FixedSizeBinaryArray>;
|
||||
|
||||
// These used to be member variables, hence the "m_" prefix. But they are
|
||||
// all constants, so I removed them as member variables. May change the
|
||||
// variable name later (to remove the m_ prefix) as time allows.
|
||||
@ -770,7 +771,7 @@ int Dctnry::insertDctnry1(Signature& curSig, bool found, char* pOut, int& outOff
|
||||
}
|
||||
else
|
||||
{
|
||||
const char* start = (const char*) curSig.signature;
|
||||
const char* start = (const char*)curSig.signature;
|
||||
const char* end = (const char*)(curSig.signature + curSig.size);
|
||||
size_t numChars = cs->numchars(start, end);
|
||||
size_t maxCharLength = m_colWidth / cs->mbmaxlen;
|
||||
@ -784,7 +785,7 @@ int Dctnry::insertDctnry1(Signature& curSig, bool found, char* pOut, int& outOff
|
||||
}
|
||||
}
|
||||
}
|
||||
else // cs->mbmaxlen == 1
|
||||
else // cs->mbmaxlen == 1
|
||||
{
|
||||
if (curSig.size > m_colWidth)
|
||||
{
|
||||
@ -971,9 +972,8 @@ int Dctnry::insertDctnry1(Signature& curSig, bool found, char* pOut, int& outOff
|
||||
* success - successfully write the header to block
|
||||
* failure - it did not write the header to block
|
||||
******************************************************************************/
|
||||
int Dctnry::insertDctnryParquet(std::shared_ptr<arrow::Array> columnData, int startRowIdx,
|
||||
const int totalRow, const int col, char* tokenBuf,
|
||||
long long& truncCount, const CHARSET_INFO* cs,
|
||||
int Dctnry::insertDctnryParquet(std::shared_ptr<arrow::Array> columnData, int startRowIdx, const int totalRow,
|
||||
const int col, char* tokenBuf, long long& truncCount, const CHARSET_INFO* cs,
|
||||
const WriteEngine::ColType& weType)
|
||||
{
|
||||
#ifdef PROFILE
|
||||
@ -993,9 +993,8 @@ int Dctnry::insertDctnryParquet(std::shared_ptr<arrow::Array> columnData, int st
|
||||
cb.file.pFile = m_dFile;
|
||||
WriteEngine::Token nullToken;
|
||||
|
||||
bool isNonNullArray = true;
|
||||
std::shared_ptr<arrow::BinaryArray> binaryArray;
|
||||
std::shared_ptr<arrow::FixedSizeBinaryArray> fixedSizeBinaryArray;
|
||||
BinaryArraySharedPtr binaryArray;
|
||||
FixedSizeBinaryArraySharedPtr fixedSizeBinaryArray;
|
||||
|
||||
if (columnData->type_id() != arrow::Type::type::FIXED_SIZE_BINARY)
|
||||
binaryArray = std::static_pointer_cast<arrow::BinaryArray>(columnData);
|
||||
@ -1003,8 +1002,7 @@ int Dctnry::insertDctnryParquet(std::shared_ptr<arrow::Array> columnData, int st
|
||||
fixedSizeBinaryArray = std::static_pointer_cast<arrow::FixedSizeBinaryArray>(columnData);
|
||||
|
||||
// check if this column data imported is NULL array or not
|
||||
if (columnData->type_id() == arrow::Type::type::NA)
|
||||
isNonNullArray = false;
|
||||
bool isNonNullArray = columnData->type_id() == arrow::Type::type::NA ? false : true;
|
||||
|
||||
//...Loop through all the rows for the specified column
|
||||
while (startPos < totalRow)
|
||||
@ -1092,8 +1090,8 @@ int Dctnry::insertDctnryParquet(std::shared_ptr<arrow::Array> columnData, int st
|
||||
}
|
||||
}
|
||||
|
||||
RETURN_ON_ERROR(insertDctnry1(curSig, found, pOut, outOffset, startPos, totalUseSize, cb, next, truncCount,
|
||||
cs, weType));
|
||||
RETURN_ON_ERROR(insertDctnry1(curSig, found, pOut, outOffset, startPos, totalUseSize, cb, next,
|
||||
truncCount, cs, weType));
|
||||
}
|
||||
|
||||
#ifdef PROFILE
|
||||
@ -1125,8 +1123,7 @@ int Dctnry::insertDctnryParquet(std::shared_ptr<arrow::Array> columnData, int st
|
||||
* failure - it did not write the header to block
|
||||
******************************************************************************/
|
||||
int Dctnry::insertDctnry(const char* buf, ColPosPair** pos, const int totalRow, const int col, char* tokenBuf,
|
||||
long long& truncCount, const CHARSET_INFO* cs,
|
||||
const WriteEngine::ColType& weType)
|
||||
long long& truncCount, const CHARSET_INFO* cs, const WriteEngine::ColType& weType)
|
||||
{
|
||||
#ifdef PROFILE
|
||||
Stats::startParseEvent(WE_STATS_PARSE_DCT);
|
||||
@ -1199,9 +1196,9 @@ int Dctnry::insertDctnry(const char* buf, ColPosPair** pos, const int totalRow,
|
||||
curSig.signature = (unsigned char*)pIn;
|
||||
}
|
||||
|
||||
RETURN_ON_ERROR(insertDctnry1(curSig, found, pOut, outOffset, startPos, totalUseSize, cb, next, truncCount,
|
||||
cs, weType));
|
||||
} // end while
|
||||
RETURN_ON_ERROR(insertDctnry1(curSig, found, pOut, outOffset, startPos, totalUseSize, cb, next,
|
||||
truncCount, cs, weType));
|
||||
} // end while
|
||||
|
||||
#ifdef PROFILE
|
||||
Stats::stopParseEvent(WE_STATS_PARSE_DCT);
|
||||
|
@ -37,11 +37,8 @@
|
||||
#include "bytestream.h"
|
||||
#include "nullstring.h"
|
||||
|
||||
|
||||
|
||||
#define EXPORT
|
||||
|
||||
|
||||
namespace arrow
|
||||
{
|
||||
class Array;
|
||||
|
Reference in New Issue
Block a user