Hi,
I have found the same problem. It seems that it happens when the fetch size
is greater than or equals to the number of returned rows. The problem is
that the first row of each page is the last row of the previous page. It
can also be easily reproduced with cqlsh:
CREATE KEYSPACE test with replication = {'class' : 'SimpleStrategy',
'replication_factor' : '1' };
USE test;
CREATE TABLE test (rc int, pk int, PRIMARY KEY ((pk)));
INSERT INTO test (pk, rc) VALUES (1, 1);
INSERT INTO test (pk, rc) VALUES (2, 2);
INSERT INTO test (pk, rc) VALUES (3, 3);
INSERT INTO test (pk, rc) VALUES (4, 4);
INSERT INTO test (pk, rc) VALUES (5, 5);
INSERT INTO test (pk, rc) VALUES (6, 6);
PAGING 4;
SELECT * FROM test WHERE pk IN (1, 2, 3); -- Ok
PAGING 3;
SELECT * FROM test WHERE pk IN (1, 2, 3); -- Wrong
PAGING 2;
SELECT * FROM test WHERE pk IN (1, 2, 3); -- Wrong
The output of the last query is:
pk | rc
----+----
1 | 1
2 | 2
---MORE---
pk | rc
----+----
2 | 2
3 | 3
---MORE---
pk | rc
----+----
3 | 3
Regards,
2016-02-22 8:54 GMT+00:00 Benjamin Lerer <benjamin.le...@datastax.com>:
> Thanks for reporting the problem.
> I will look into it.
>
> Regards,
>
> Benjamin
>
> On Sun, Feb 21, 2016 at 3:19 PM, Luis Ángel Vicente Sánchez <
> langel.gro...@gmail.com> wrote:
>
> > I think I have found an issue when you have a simple column family with a
> > simple primary key (no clustering keys) and you use the IN clause. This
> > problem only happens with Cassandra 3.x, in Cassandra 2.x works just
> fine.
> >
> > You can reproduce this issue by checking out this project:
> >
> > https://github.com/lvicentesanchez/cassandra-3.x-error
> >
> > It's written in Scala but I have provided instructions on how to run the
> > program.
> >
> > I will copy here the content of README.md where I explain the problem:
> >
> > # cassandra-3.x-error #
> >
> > This repository reproduces an error in Cassandra 3.x when the fetch size
> is
> > smaller than the number of rows in a column family. For the shake of this
> > demo, fetch size would be 2 and the number of rows 3.
> >
> > ## Pre-steps ##
> >
> > 1. Use `src/main/cql/script.cql` to create the key space and column
> > families.
> > 2. Modify contact point and port in
> `src/main/scala/CassandraError.scala`.
> > 3. Modify contact point and port in `src/main/scala/CassandraOk.scala`.
> > 4. Open `sbt` by executing `./sbt` in the root of the project.
> >
> > ## Error ##
> >
> > Execute `runMain CassandraError`. We would get this in our terminal:
> >
> > ```
> > [info] Measure(1,11)
> > [info] Measure(2,7)
> > [info] Measure(2,7)
> > [info] Measure(3,23)
> > [info] Measure(3,23)
> > ```
> >
> > If we run the same program against Cassandra 2.x, we would get this:
> >
> > ```
> > [info] Measure(1,11)
> > [info] Measure(2,7)
> > [info] Measure(3,23)
> > ```
> >
> > We were expecting Cassandra 3.x to produce the same output as Cassandra
> > 2.x. It seems that there is an issue with simple primary keys, pagination
> > and the IN clause.
> >
> > ## Ok ##
> >
> > Execute `runMain CassandraOk`. Whether we ran it against Cassandra 2.x or
> > Cassandra 3.x, we would get this in our terminal:
> >
> > ```
> > [info] Measure(1,0,11)
> > [info] Measure(2,0,7)
> > [info] Measure(3,0,23)
> > ```
> >
> > Should I create an issue?
> >
> > Regards,
> >
> > Luis
> >
>
--
Andrés de la Peña
Vía de las dos Castillas, 33, Ática 4, 3ª Planta
28224 Pozuelo de Alarcón, Madrid
Tel: +34 91 828 6473 // www.stratio.com // *@stratiobd
<https://twitter.com/StratioBD>*
