void CDiskWriteSlaveActivityBase::open()
{
if (dlfn.isExternal() && !firstNode())
{
input.setown(createDataLinkSmartBuffer(this, inputs.item(0), PROCESS_SMART_BUFFER_SIZE, isSmartBufferSpillNeeded(this), grouped, RCUNBOUND, NULL, false, &container.queryJob().queryIDiskUsage()));
startInput(input);
if (!rfsQueryParallel)
{
ActPrintLog("Blocked, waiting for previous part to complete write");
CMessageBuffer msg;
if (!receiveMsg(msg, container.queryJob().queryMyRank()-1, mpTag))
return;
rowcount_t prevRows;
msg.read(prevRows);
msg.read(tempExternalName); // reuse temp filename, last node will rename
ActPrintLog("Previous write row count = %"RCPF"d", prevRows);
}
}
else
{
input.set(inputs.item(0));
startInput(input);
}
processed = THORDATALINK_STARTED;
bool extend = 0 != (diskHelperBase->getFlags() & TDWextend);
if (extend)
ActPrintLog("Extending file %s", fName.get());
size32_t exclsz = 0;
calcFileCrc = true;
bool external = dlfn.isExternal();
bool query = dlfn.isQuery();
if (query && compress)
UNIMPLEMENTED;
bool direct = query || (external && !firstNode());
bool rename = !external || (!query && lastNode());
Owned<IFileIO> iFileIO = createMultipleWrite(this, *partDesc, exclsz, compress, extend||(external&&!query), ecomp, this, direct, rename, &abortSoon, (external&&!query) ? &tempExternalName : NULL);
if (compress)
{
ActPrintLog("Performing row compression on output file: %s", fName.get());
calcFileCrc = false;
}
Owned<IFileIOStream> stream;
if (wantRaw())
{
outraw.setown(createBufferedIOStream(iFileIO));
stream.set(outraw);
}
else
{
stream.setown(createIOStream(iFileIO));
out.setown(createRowWriter(stream,::queryRowSerializer(input),::queryRowAllocator(input),grouped,calcFileCrc,false)); // flushed by close
}
CDfsLogicalFileName dlfn;
dlfn.set(logicalFilename);
if (extend || (external && !query))
stream->seek(0,IFSend);
ActPrintLog("Created output stream for %s", fName.get());
}
void CDiskWriteSlaveActivityBase::open()
{
if (dlfn.isExternal() && !firstNode())
{
input.setown(createDataLinkSmartBuffer(this, inputs.item(0), PROCESS_SMART_BUFFER_SIZE, isSmartBufferSpillNeeded(this), grouped, RCUNBOUND, NULL, false, &container.queryJob().queryIDiskUsage()));
startInput(input);
if (!rfsQueryParallel)
{
ActPrintLog("Blocked, waiting for previous part to complete write");
CMessageBuffer msg;
if (!receiveMsg(msg, queryJobChannel().queryMyRank()-1, mpTag))
return;
rowcount_t prevRows;
msg.read(prevRows);
msg.read(tempExternalName); // reuse temp filename, last node will rename
ActPrintLog("Previous write row count = %" RCPF "d", prevRows);
}
}
else
{
input.set(inputs.item(0));
startInput(input);
}
processed = THORDATALINK_STARTED;
bool extend = 0 != (diskHelperBase->getFlags() & TDWextend);
if (extend)
ActPrintLog("Extending file %s", fName.get());
/* Fixed length record size is used when outputting compressed stream to determine run-length compression vs default LZW compression.
* NB: only for FLAT files, not CSV or XML
*/
size32_t diskRowMinSz = 0;
IOutputMetaData *diskRowMeta = diskHelperBase->queryDiskRecordSize()->querySerializedDiskMeta();
if (diskRowMeta->isFixedSize() && (TAKdiskwrite == container.getKind()))
{
diskRowMinSz = diskRowMeta->getMinRecordSize();
if (grouped)
diskRowMinSz += 1;
}
calcFileCrc = true;
bool external = dlfn.isExternal();
bool query = dlfn.isQuery();
if (query && compress)
UNIMPLEMENTED;
unsigned twFlags = external ? TW_External : 0;
if (query || (external && !firstNode()))
twFlags |= TW_Direct;
if (!external || (!query && lastNode()))
twFlags |= TW_RenameToPrimary;
if (extend||(external&&!query))
twFlags |= TW_Extend;
Owned<IFileIO> iFileIO = createMultipleWrite(this, *partDesc, diskRowMinSz, twFlags, compress, ecomp, this, &abortSoon, (external&&!query) ? &tempExternalName : NULL);
if (compress)
{
ActPrintLog("Performing row compression on output file: %s", fName.get());
// NB: block compressed output has implicit crc of 0, no need to calculate in row writer.
calcFileCrc = false;
}
Owned<IFileIOStream> stream;
if (wantRaw())
{
outraw.setown(createBufferedIOStream(iFileIO));
stream.set(outraw);
}
else
{
stream.setown(createIOStream(iFileIO));
unsigned rwFlags = 0;
if (grouped)
rwFlags |= rw_grouped;
if (calcFileCrc)
rwFlags |= rw_crc;
out.setown(createRowWriter(stream, ::queryRowInterfaces(input), rwFlags));
}
if (extend || (external && !query))
stream->seek(0,IFSend);
ActPrintLog("Created output stream for %s", fName.get());
}
void start()
{
ActivityTimer s(totalCycles, timeActivities, NULL);
count = 0;
eof = nextPut = false;
inrowif.set(::queryRowInterfaces(inputs.item(0)));
if (global) // only want lookahead if global (hence serial)
input.setown(createDataLinkSmartBuffer(this, inputs.item(0),ITERATE_SMART_BUFFER_SIZE,isSmartBufferSpillNeeded(this),false,RCUNBOUND,NULL,false,&container.queryJob().queryIDiskUsage())); // only allow spill if input can stall
else
input.set(inputs.item(0));
try
{
startInput(input);
}
catch (IException *e)
{
ActPrintLog(e,"ITERATE");
throw;
}
dataLinkStart("ITERATOR", container.queryId());
}
virtual void setInputStream(unsigned index, CThorInput &_input, bool consumerOrdered) override
{
PARENT::setInputStream(index, _input, consumerOrdered);
if (!isFastThrough(input))
setLookAhead(0, createRowStreamLookAhead(this, inputStream, queryRowInterfaces(input), CHOOSESETS_SMART_BUFFER_SIZE, isSmartBufferSpillNeeded(this), false, RCUNBOUND, NULL, &container.queryJob().queryIDiskUsage()));
}
void CDiskWriteSlaveActivityBase::setInputStream(unsigned index, CThorInput &_input, bool consumerOrdered)
{
PARENT::setInputStream(index, _input, consumerOrdered);
if (dlfn.isExternal() && !firstNode())
setLookAhead(0, createRowStreamLookAhead(this, inputStream, queryRowInterfaces(input), PROCESS_SMART_BUFFER_SIZE, isSmartBufferSpillNeeded(this), grouped, RCUNBOUND, NULL, &container.queryJob().queryIDiskUsage()));
}
