CODEWITHSUNDEEP
paginationcassandra
sribalaji a
sribalaji a

Reputation: 1

Cassandra: How to paginate back from a 'n' th row key using RangeSlicesQuery?

How to paginate back from a 'n' th row key using RangeSlicesQuery ?

The data is based on RandomPartitioner and the reverse option on the setRange function seems to be applicable only for the columns not on the rows. (http://tinyurl.com/73tncn3)

Example :

row_key1

row_key2

row_key3

row_key4

row_key5

row_key6

Expected :

row_key3

row_key4

Recieved :

row_key1

row_key2

Received

row_key4

row_key5

Upvotes: 0

Views: 537

Answers (1)

tom.wilkie
tom.wilkie

Reputation: 2866

When using the RandomPartitioner, the effective order in which Cassandra store the rows is random. Therefore you cannot do range slices in the way you expect.

What you should do is materialise your view in a row, and use some sort of calculated bucket for the row key. For example, the row key could be the the real row key / 10, and then the row could contain 10 real rows:

0 -> {row_key1 : v1, row_key2: v2, ...}
1 -> {row_key10: v10, row_key11: v11, ...}
and do on.

To do your query now, you do a get slice on row 0, from columns 'row_key4'...

Now you may ask, "what about my column keys"? For this you should use compound column names. For example, something that originally looked like this:

row_key1 -> {column1: v1, column2: v2, ...}

Would now look like this:

0 -> {[row_key1, column1]: v1, [row_key2, column2]: v2, ...}
and so on.

HTH

Tom

Upvotes: 1

Related Questions