Please consider this thread closed.

This was a sloppy description of the bug, written by me at a time when I should actually be sleeping.

I wrote a more detailed specification at a later time, and the result of it was that the bug was confirmed and in fact it had already been solved (see CASSANDRA-703).

I used cassandra 0.5 when observing this bug.

Sorry for any inconvenience. I shouldn't have written this sloppy descr. :-/


On Mon, Apr 5, 2010 at 12:06 AM, Eric Evans <> wrote:
On Sat, 2010-04-03 at 02:55 +0200, Arash Bazrafshan wrote:
> Think I got a bug in Cassandra. Do you also think it's a bug?
> It should be noted that I experience this bug when using cassandra
> through thrift's php api (the low-level one generated by thrift, not
> some high-level from the cassandra wiki).
> It happens when i insert a supercolumn (with a single column "bar")
> under key "foo", and then delete the supercolumn (under the same key).
> After that if I insert the same supercolumn (with the same column)
> under key "foo" then I'm not able to retrieve it (the entire
> supercolumn).
> However I am able to retrieve the single column "bar" within the
> supercolumn.
> So If I try to retrieve the entire supercolumn then I get
> "notfoundexception", but if I wanna fetch the single column "bar"
> within the supercolumn, then that's fine.

See if you can replicate this using the bin/cassandra-cli. I've attached
a file with the equivalent commands, you can pipe it to stdin on the
command line (i.e. bin/cassandra-cli --host localhost <


FWIW, I wasn't able to replicate this (

Eric Evans