cql3 - Do tables share partitions ("low level rows") in Cassandra? -

Assume that I have two tables defined by the CLL (column family).

  create table one (PKUUID, CKAAT, well text, primary key (PK, CKA)); Make Table B (PKUUID, CKB Text, Val 1 Boolean, Wall 2 Decimal, Primary K (PK, CKB));  

If I now include one line in each table with the same partition key:

  INSERT (pk, cka, val) VALUES (' F47ac10b) -58cc-4372-a567-0e02b2a337'9 ', 5,' hi '); INSERT ('f47ac10b-58cc-4372-a567-0e02b2a3d379', 'x', 'hello', 'hey') in b (pk, ckb, val 1, val 2) values;  

Will there now be 1 or 2 rows at the storage level?

There will be 2.

The data in Cassandra is written in "memtables", and then the "SSTables" on-disk drive is kept both memable and SSTable based on a per-column family , So the rows in different column families (tables) always make different rows at the storage level.

Cassandra writes before a committed log (for durability), and then called a memory memory structure called memtable. After writing one it succeeds that it is written on logs logs and memory, so there is very little disk I / O at the time of writing. I is batch in written memory and from time to time a permanent table structure Which is called SSTAB (sorted string table). Memtables and SSTables per column family is maintained. Memtables are arranged in sequence order by row keys and flushed to SSTables sequentially (in the form of a random relational database).