EJTTianyu created IOTDB-401:
-------------------------------

             Summary: Flushed data size inconsistent exception when writing 
measurement(same name but different device) with different data type
                 Key: IOTDB-401
                 URL: https://issues.apache.org/jira/browse/IOTDB-401
             Project: Apache IoTDB
          Issue Type: Bug
            Reporter: EJTTianyu


Hi,
 My current project requires exporting data from Kairosdb as tsfile. I use the 
following steps to complete this task.
 1. Query and export kairosdb data to csv;
 2. Export csv to tsfile.

In the second step, I encountered the following exception:
{code:java}
ERROR org.apache.iotdb.tsfile.write.chunk.ChunkWriterImpl - meet error in 
chunkBuffer.writePageHeaderAndDataIntoBuff, ignore this page:
org.apache.iotdb.tsfile.exception.write.PageException: No valid data point in 
this page
        at 
org.apache.iotdb.tsfile.write.chunk.ChunkBuffer.writePageHeaderAndDataIntoBuff(ChunkBuffer.java:91)
        at 
org.apache.iotdb.tsfile.write.chunk.ChunkWriterImpl.writePage(ChunkWriterImpl.java:308)
        at 
org.apache.iotdb.tsfile.write.chunk.ChunkWriterImpl.sealCurrentPage(ChunkWriterImpl.java:351)
        at 
org.apache.iotdb.tsfile.write.chunk.ChunkGroupWriterImpl.sealAllChunks(ChunkGroupWriterImpl.java:155)
        at 
org.apache.iotdb.tsfile.write.chunk.ChunkGroupWriterImpl.flushToFileWriter(ChunkGroupWriterImpl.java:124)
        at 
org.apache.iotdb.tsfile.write.TsFileWriter.flushAllChunkGroups(TsFileWriter.java:310)
        at 
org.apache.iotdb.tsfile.write.TsFileWriter.close(TsFileWriter.java:337)

ERROR kairosdb.export.csv.TransToTsfile - error occurs when generate tsfile
java.io.IOException: Flushed data size is inconsistent with computation! 
Estimated: 29, Actual: 0
        at 
org.apache.iotdb.tsfile.write.TsFileWriter.flushAllChunkGroups(TsFileWriter.java:312)
        at 
org.apache.iotdb.tsfile.write.TsFileWriter.close(TsFileWriter.java:337)
{code}
You can use the following strategy to reproduce this bug. 
 First create a time series like d0.s0, whose data type is INT64. Write some 
data point to the TsfileWriter. Second create a time series like d1.s0, whose 
data type is FLOAT. When writing TSRecord to the same tsfile writer, the bug 
occurs.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to