pavibhai commented on a change in pull request #668:
URL: https://github.com/apache/orc/pull/668#discussion_r643255303
##########
File path: java/core/src/java/org/apache/orc/impl/reader/StripePlanner.java
##########
@@ -445,55 +509,83 @@ private static boolean hasSomeRowGroups(boolean[]
includedRowGroups) {
* @return the list of disk ranges that will be loaded
*/
private BufferChunkList planPartialDataReading(OrcIndex index,
- @NotNull boolean[]
includedRowGroups) {
+ @NotNull boolean[]
includedRowGroups,
+ int startGroup,
+ TypeReader.ReadPhase
readPhase) {
BufferChunkList result = new BufferChunkList();
if (hasSomeRowGroups(includedRowGroups)) {
InStream.StreamOptions compression = dataReader.getCompressionOptions();
boolean isCompressed = compression.getCodec() != null;
int bufferSize = compression.getBufferSize();
OrcProto.RowIndex[] rowIndex = index.getRowGroupIndex();
for (StreamInformation stream : dataStreams) {
- if (RecordReaderUtils.isDictionary(stream.kind,
encodings[stream.column])) {
- addChunk(result, stream, stream.offset, stream.length);
+ if (readPhase == TypeReader.ReadPhase.ALL
+ ||(readPhase == TypeReader.ReadPhase.LEADERS
Review comment:
> nit. space after ||
done
> nit. lets keep conditions on a single line
brought the subcondition into a single line not the entire condition as it
extends the line boundary significantly
> nit. reduce new lines (e.g., processStream call)
Fixed the processStream call however I am unclear on how to handle
parameters. I have generally been following chompdown if long, I see both
variants in the code chomp or just wrap.
--
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]
