pgaref commented on a change in pull request #648:
URL: https://github.com/apache/orc/pull/648#discussion_r586348582
##########
File path: c++/src/Reader.cc
##########
@@ -892,29 +939,67 @@ namespace orc {
void RowReaderImpl::startNextStripe() {
reader.reset(); // ColumnReaders use lots of memory; free old memory first
- currentStripeInfo = footer->stripes(static_cast<int>(currentStripe));
- uint64_t fileLength = contents->stream->getLength();
- if (currentStripeInfo.offset() + currentStripeInfo.indexlength() +
+ rowIndexes.clear();
+ bloomFilterIndex.clear();
+
+ do {
+ currentStripeInfo = footer->stripes(static_cast<int>(currentStripe));
+ uint64_t fileLength = contents->stream->getLength();
+ if (currentStripeInfo.offset() + currentStripeInfo.indexlength() +
currentStripeInfo.datalength() + currentStripeInfo.footerlength() >=
fileLength) {
- std::stringstream msg;
- msg << "Malformed StripeInformation at stripe index " << currentStripe
<< ": fileLength="
- << fileLength << ", StripeInfo=(offset=" <<
currentStripeInfo.offset() << ", indexLength="
- << currentStripeInfo.indexlength() << ", dataLength=" <<
currentStripeInfo.datalength()
- << ", footerLength=" << currentStripeInfo.footerlength() << ")";
- throw ParseError(msg.str());
+ std::stringstream msg;
+ msg << "Malformed StripeInformation at stripe index " << currentStripe
<< ": fileLength="
+ << fileLength << ", StripeInfo=(offset=" <<
currentStripeInfo.offset() << ", indexLength="
+ << currentStripeInfo.indexlength() << ", dataLength=" <<
currentStripeInfo.datalength()
+ << ", footerLength=" << currentStripeInfo.footerlength() << ")";
+ throw ParseError(msg.str());
+ }
+ currentStripeFooter = getStripeFooter(currentStripeInfo,
*contents.get());
+ rowsInCurrentStripe = currentStripeInfo.numberofrows();
+
+ if (sargsApplier) {
+ // read row group statistics and bloom filters of current stripe
+ loadStripeIndex();
+
+ // select row groups to read in the current stripe
+ sargsApplier->pickRowGroups(rowsInCurrentStripe,
+ rowIndexes,
+ bloomFilterIndex);
+ if (sargsApplier->hasSelectedFrom(currentRowInStripe)) {
+ // current stripe has at least one row group matching the predicate
+ break;
+ } else {
+ // advance to next stripe when current stripe has no matching rows
+ currentStripe += 1;
+ currentRowInStripe = 0;
+ }
+ }
+ } while (sargsApplier && currentStripe < lastStripe);
+
+ if (currentStripe < lastStripe) {
+ const Timezone& writerTimezone =
Review comment:
Lets add some doc here
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
