[
https://issues.apache.org/jira/browse/DRILL-5546?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16135707#comment-16135707
]
ASF GitHub Bot commented on DRILL-5546:
---------------------------------------
Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/906#discussion_r134301372
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/ScanBatch.java
---
@@ -152,97 +157,75 @@ public void kill(boolean sendUpstream) {
}
}
- private void releaseAssets() {
- container.zeroVectors();
- }
-
- private void clearFieldVectorMap() {
- for (final ValueVector v : mutator.fieldVectorMap().values()) {
- v.clear();
- }
- }
-
@Override
public IterOutcome next() {
if (done) {
return IterOutcome.NONE;
}
oContext.getStats().startProcessing();
try {
- try {
- injector.injectChecked(context.getExecutionControls(),
"next-allocate", OutOfMemoryException.class);
-
- currentReader.allocate(mutator.fieldVectorMap());
- } catch (OutOfMemoryException e) {
- clearFieldVectorMap();
- throw UserException.memoryError(e).build(logger);
- }
- while ((recordCount = currentReader.next()) == 0) {
+ while (true) {
try {
- if (!readers.hasNext()) {
- // We're on the last reader, and it has no (more) rows.
- currentReader.close();
- releaseAssets();
- done = true; // have any future call to next() return NONE
-
- if (mutator.isNewSchema()) {
- // This last reader has a new schema (e.g., we have a
zero-row
- // file or other source). (Note that some sources have a
non-
- // null/non-trivial schema even when there are no rows.)
+ injector.injectChecked(context.getExecutionControls(),
"next-allocate", OutOfMemoryException.class);
+ currentReader.allocate(mutator.fieldVectorMap());
+ } catch (OutOfMemoryException e) {
+ clearFieldVectorMap();
+ throw UserException.memoryError(e).build(logger);
+ }
- container.buildSchema(SelectionVectorMode.NONE);
- schema = container.getSchema();
+ recordCount = currentReader.next();
+ Preconditions.checkArgument(recordCount >= 0,
+ "recordCount from RecordReader.next() should not be negative");
- return IterOutcome.OK_NEW_SCHEMA;
- }
- return IterOutcome.NONE;
- }
- // At this point, the reader that hit its end is not the last
reader.
+ boolean isNewRegularSchema = mutator.isNewSchema();
+ // We should skip the reader, when recordCount = 0 && !
isNewRegularSchema.
+ // Add/set implicit column vectors, only when reader gets > 0 row,
or
+ // when reader gets 0 row but with a schema with new field added
+ if (recordCount > 0 || isNewRegularSchema) {
+ addImplicitVectors();
+ populateImplicitVectors();
+ }
- // If all the files we have read so far are just empty, the
schema is not useful
- if (! hasReadNonEmptyFile) {
- container.clear();
- clearFieldVectorMap();
- mutator.clear();
- }
+ boolean isNewImplicitSchema = mutator.isNewSchema();
+ for (VectorWrapper<?> w : container) {
+ w.getValueVector().getMutator().setValueCount(recordCount);
+ }
+ final boolean isNewSchema = isNewRegularSchema ||
isNewImplicitSchema;
+ oContext.getStats().batchReceived(0, recordCount, isNewSchema);
+ if (recordCount == 0) {
currentReader.close();
- currentReader = readers.next();
- implicitValues = implicitColumns.hasNext() ?
implicitColumns.next() : null;
- currentReader.setup(oContext, mutator);
- try {
- currentReader.allocate(mutator.fieldVectorMap());
- } catch (OutOfMemoryException e) {
- clearFieldVectorMap();
- throw UserException.memoryError(e).build(logger);
+ if (isNewSchema) {
+ // current reader presents a new schema in mutator even though
it has 0 row.
--- End diff --
Thanks for the comments here and in the PR. I wonder, can the PR
description be moved into a class Javadoc comment so that it is available for
future readers? Also, makes it easier to review since the description is close
to the code.
> Schema change problems caused by empty batch
> --------------------------------------------
>
> Key: DRILL-5546
> URL: https://issues.apache.org/jira/browse/DRILL-5546
> Project: Apache Drill
> Issue Type: Bug
> Reporter: Jinfeng Ni
> Assignee: Jinfeng Ni
>
> There have been a few JIRAs opened related to schema change failure caused by
> empty batch. This JIRA is opened as an umbrella for all those related JIRAS (
> such as DRILL-4686, DRILL-4734, DRILL4476, DRILL-4255, etc).
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
