Dear Jeremy Dunck,

I tried to compact, and get and error:

Caused by: java.io.UTFDataFormatException: malformed input around byte 13
    at java.io.DataInputStream.readUTF(DataInputStream.java:617)
    at java.io.RandomAccessFile.readUTF(RandomAccessFile.java:887)
    at org.apache.cassandra.io.IteratingRow.<init>(IteratingRow.java:48)
    at org.apache.cassandra.io.SSTableScanner$KeyScanningIterator.next(SSTableScanner.java:115)

I also tried in version 0.6, but above error still exist.

Perhaps, I will tried the way David Timothy suggest.

@Stu Hood: Do you implement code for 808 issues?

Thank a lot for support.

On Wed, Mar 31, 2010 at 11:02 AM, Jeremy Dunck <jdunck@gmail.com> wrote:
On Wed, Mar 31, 2010 at 7:55 AM, Stu Hood <stu.hood@rackspace.com> wrote:
> Eventually the new file format will make it in with #674, and we'll be able to implement an option to skip corrupted data:
>
> https://issues.apache.org/jira/browse/CASSANDRA-808

That ticket seems to indicate that compaction will remove the corrupt
data, and might make the later data readable.

JKnight, did you try compacting?



--
Best regards,
JKnight