[
https://issues.apache.org/jira/browse/DRILL-7053?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16776460#comment-16776460
]
Paul Rogers commented on DRILL-7053:
------------------------------------
[~vitalii], thanks for taking a look at this bug. It is interesting that you
did not see the issue. Here's what I did:
{noformat}
cd drill
mvn checkout master
mvn clean install -DskipTests
{noformat}
Then, in Eclipse:
* Open the
[org.apache.drill.exec.store.easy.text.compliant.TestCsv.java](https://github.com/apache/drill/blob/master/exec/java-exec/src/test/java/org/apache/drill/exec/store/easy/text/compliant/TestCsv.java)
test case.
* Run the test with "Debug As... --> JUnit Test".
Turns out, the test that issues the error is this one, which is designed to
test an error case:
{code:java}
@Test
public void testEmptyCsvHeaders() throws IOException {
String fileName = "case1.csv";
buildFile(fileName, emptyHeaders);
try {
client.queryBuilder().sql(makeStatement(fileName)).run();
fail();
} catch (Exception e) {
assertTrue(e.getMessage().contains("must define at least one header"));
}
}
{code}
So, the problem likely has something to do with cleaning up a failed fragment
execution. Maybe only in debug mode (assertions enabled)?
> Benign, but unexpected, failure in CsvTest
> ------------------------------------------
>
> Key: DRILL-7053
> URL: https://issues.apache.org/jira/browse/DRILL-7053
> Project: Apache Drill
> Issue Type: Bug
> Affects Versions: 1.15.0
> Reporter: Paul Rogers
> Priority: Minor
>
> Run the unit test {{CsvTest}} from the {{master}} branch in the Eclipse
> debugger. The test will emit the following to the console:
> {noformat}
> 17:54:58.347 [238e068f-7d4e-838b-7e9e-b6958f977f65:frag:0:0] ERROR
> o.a.d.e.physical.impl.BaseRootExec - Batch dump started: dumping last 2
> failed batches
> 17:54:58.351 [238e068f-7d4e-838b-7e9e-b6958f977f65:frag:0:0] ERROR
> o.a.d.e.p.i.v.IteratorValidatorBatchIterator -
> IteratorValidatorBatchIterator[container=org.apache.drill.exec.record.VectorContainer@549d2d70[recordCount
> = 0, schemaChanged = true, schema = null, wrappers = [], ...], instNum=3,
> batchTypeName=ScanBatch, lastSchema=null, lastNewSchema=null]
> 17:54:58.351 [238e068f-7d4e-838b-7e9e-b6958f977f65:frag:0:0] ERROR
> o.a.d.exec.physical.impl.ScanBatch -
> ScanBatch[container=org.apache.drill.exec.record.VectorContainer@549d2d70[recordCount
> = 0, schemaChanged = true, schema = null, wrappers = [], ...],
> currentReader=CompliantTextRecordReader[File=file:/Users/paulrogers/git/drill/exec/java-exec/target/org.apache.drill.exec.store.easy.text.compliant.TestCsv/data/case1.csv,
> reader=TextReader[Line=1, Column=2, Record=1, Byte pos=2]], schema=null]
> 17:54:58.351 [238e068f-7d4e-838b-7e9e-b6958f977f65:frag:0:0] ERROR
> o.a.d.e.physical.impl.BaseRootExec - Batch dump completed.
> {noformat}
> The tests all pass. It is not clear why we get the above error. However, it
> is something that should be investigated and fixed since it causes "noise"
> when trying to get a clean test run.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
