Let say I have 6 nodes with replication factor of 3.
Column Family: table1
Clustering key: C1, C2
Sample Data (9 rows)
COL1 | COL2 | COL3 a1 | 01 | 'Hello' a1 | 02 | 'Hi' a1 | 03 | 'Good Morning' b1 | 01 | 'Good Afternoon' b1 | 02 | 'Lunch Time' c1 | 01 | 'Night time' c1 | 02 | 'Good night' c1 | 03 | 'Bed time' c1 | 04 | 'Tired'
is there a way to make all of "a1" to stored together in one node so that the coordinator does need to go across the cluster?
Yes, there is. If column 1 is the partition key of your table and col2 is the clustering key then all rows will be stored in the same partition.
The schema for the table will look something like this:
CREATE TABLE table1
col1 text,
col2 text,
col2 text,
PRIMARY KEY ((col1), col2)
)
If you're not familiar with partitions and clustering columns, I've explained them in a bit more detail in question #6171.
It also seems like you might be interested in learning more about how data is partitioned in Cassandra so I highly recommend you have a look at the short free interactive tutorials at https://www.datastax.com/learn/data-modeling-by-example. Cheers!
6 People are following this question.
