You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "miaojianlong (Jira)" <ji...@apache.org> on 2021/12/29 16:09:00 UTC
[jira] [Created] (FLINK-25482) Hive Lookup Join with decimal type failed
miaojianlong created FLINK-25482:
------------------------------------
Summary: Hive Lookup Join with decimal type failed
Key: FLINK-25482
URL: https://issues.apache.org/jira/browse/FLINK-25482
Project: Flink
Issue Type: Bug
Components: Connectors / Hive
Affects Versions: 1.14.2, 1.13.5, 1.12.3
Reporter: miaojianlong
Hive Lookup Join with decimal type failed.
Error exception place :
{code:java}
public BigInteger(byte[] val) {
if (val.length == 0)
throw new NumberFormatException("Zero length BigInteger");{code}
Because of this code:org.apache.flink.connectors.hive.read.HiveInputFormatPartitionReader#hasNext
{code:java}
private boolean hasNext() throws IOException {
if (inputSplits.length > 0) {
if (hiveTableInputFormat.reachedEnd() && readingSplitId == inputSplits.length - 1) {
return false;
} else if (hiveTableInputFormat.reachedEnd()) {
readingSplitId++;
hiveTableInputFormat.open(inputSplits[readingSplitId]);
}
return true;
}
return false;
}{code}
when we have more than one file in hive table and when we reachedEnd of the first inputSplits.Then at this time it will open the next inputSplit,but it did not trigger the batch reading of the file.Then after calling readRecord later, the data read is empty.For decimal type, it will trigger The above exception.In other scenarios, it will read more empty row.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)