Ok, I'll look what's wrong.
--
Sylvain
On Mon, Oct 1, 2012 at 1:27 AM, Jonathan Rudenberg
<jonat...@titanous.com> wrote:
> I'm putting a collection onto the wire in an EXECUTE request with exactly the
> same bytes that were received and I get:
>
> "Can't apply operation on column with
> org.apache.cassandra.db.marshal.SetType(org.apache.cassandra.db.marshal.UTF8Type)
> type"
>
> Here's the full trace log from cassandra:
>
> DEBUG 19:24:15,414 Received: PREPARE INSERT INTO things (id, set_text) VALUES
> (?, ?);
> TRACE 19:24:15,414 CQL QUERY: INSERT INTO things (id, set_text) VALUES (?, ?);
> TRACE 19:24:15,415 Stored prepared statement #413587006 with 2 bind markers
> DEBUG 19:24:15,415 Responding: RESULT PREPARED 413587006
> [id(gocql_collections, things),
> org.apache.cassandra.db.marshal.TimeUUIDType][set_text(gocql_collections,
> things),
> org.apache.cassandra.db.marshal.SetType(org.apache.cassandra.db.marshal.UTF8Type)]
> DEBUG 19:24:15,416 Received: EXECUTE 413587006 with 2 values
> TRACE 19:24:15,416 [1] 'java.nio.HeapByteBuffer[pos=18 lim=34 cap=53]'
> TRACE 19:24:15,416 [2] 'java.nio.HeapByteBuffer[pos=38 lim=51 cap=53]'
> DEBUG 19:24:15,417 Responding: ERROR INVALID: Can't apply operation on column
> with
> org.apache.cassandra.db.marshal.SetType(org.apache.cassandra.db.marshal.UTF8Type)
> type.
>
>
> The prepared statement is:
>
> INSERT INTO things (id, set_text) VALUES (?, ?);
>
> and the collection value ({'asdf', 'sdf'}) is encoded as:
>
> 00 02 00 04 61 73 64 66 00 03 73 64 66
>
> which is byte-for-byte exactly the same as what I get from Cassandra off the
> wire when I do a query for the same data.
>
> I already have the driver working with other queries/inserts with all types,
> so this is just a collection encoding problem.
>
> Thanks,
>
> Jonathan
>
>
> On Sep 30, 2012, at 2:34 PM, Sylvain Lebresne <sylv...@datastax.com> wrote:
>
>> On Sun, Sep 30, 2012 at 6:02 PM, Jonathan Rudenberg
>> <jonat...@titanous.com> wrote:
>>> I found CASSANDRA-4453, but this only seems to specify the response format.
>>
>> It's the same format for EXECUTE.
>>
>>> Any pointers would be appreciated.
>>
>> I'm afraid there isn't much more than the source at this point but
>> I'll try to add something about that in the protocol spec.
>>
>> --
>> Sylvain
>
