Partial rows while paging.

8 views

Skip to first unread message

Robert Arritt

unread,

May 31, 2016, 11:01:27 AM5/31/16

to DataStax C++ Driver for Apache Cassandra User Mailing List

Recently I have noticed that while paging through some of our tables I am seeing partial rows returned by Cassandra. At the start of each new page I appear to get a duplicate of the previous row (although all of the clustering columns are null). My question is, is there a generic way via the driver to detect these rows (hopefully without needing to pass in knowledge of the clustering columns). If not, does anyone know if there is a way to prevent these duplicate rows from occurring?

Here are some examples. The null c2 and c3 rows only show up while I am paging

cqlsh:ks_1> select c1, c2, c3 from ks_1.table
...
media_uid_167 | |

---MORE---
c1 | c2 | c3
---------------+------+-----------
media_uid_167 | null | null

cqlsh:ks_1> select count(*) from ks_1.table_1 ;

count
-------
202

(1 rows)
cqlsh:ks_1> PAGING OFF
Disabled Query paging.
cqlsh:ks_1> select count(*) from ks_1.table_1 ;