choubenson opened a new pull request, #9702:
URL: https://github.com/apache/iotdb/pull/9702

   **Bug Description**
   1. If the compression method of the aligned value series is different from 
that of the time series, an error will be reported when decompressing.
   2. If the compression method of the aligned value series is different from 
that of the time series, the target file will be corrupted.
   
   **Reason**
   1. When reading the source file, all value columns share an uncompressor, 
resulting in decompression failure.
   2. The compression method of the time column in the alignment timeseries 
takes the first value column. If the source file is in the order of s0 s1 s2, 
and during fast compaction, the read order is s1 s0 s2 and then written to the 
target file in this order, that is , the time column of the original file is 
actually compressed according to the s0 method, and the newly written time 
column is compressed according to the s1 method, so if the compression method 
of s1 is different from s0, the target file will be corrupted if flushing time 
page or chunk directly.
   
   **Solution**
   1. When reading files, each value column needs to have its own umcompressor.
   2. For the alignment timeseries, when writing a new file, the compression 
method of the time column uses the default snappy; when rewriting the file, the 
compression method of the time column is consistent with that of the source file


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to