Database Reference
In-Depth Information
Collumn F
in Famiilly
A column family is roughly analogous to a table in a relational model. It is a container for an
ordered collection of columns.
Because each column family is stored in a separate file, be sure to define columns that you
are likely to access together in the same column family.
You define your application's column families in the Cassandra configuration file. You can
supply global (per-keyspace) values for the size of the row cache, the size of the key cache,
and the “read repair chance.” Column families can be one of two types: standard or super.
See also
Column
,
Keyspace
,
Super Column
.
Collumn N
n Name
The name part of the name/value pair stored within a
Row
.
Collumn V
n Vallue
limited by machine memory.
Commiit L
t Log
The commit log is responsible for all of the write operations in Cassandra. When you per-
form a write, it first enters the commit log so the data won't be lost in the event of failure;
then the value is populated in the memtable so it's available in memory for performance.
It is represented by the
org.apache.cassandra.db.commitlog.CommitLog
class. On
every write or delete, an entry in the form of a
RowMutation
object is serialized and appen-
ded to the commit log. These objects are organized into commit log segments. By default,
commit logs roll once they reach a size threshold of 128MB; when a new commit log is cre-
ated, it accepts writes in transit. This setting is configurable.
Compacttiion
Compaction is the process of freeing up space by merging large accumulated datafiles. This
is roughly analogous to rebuilding a table in the relational world. On compaction, the merged
data is sorted, a new index is created over the sorted data, and the freshly merged, sorted,
and indexed data is written to a single new file.
The operations that are performed during compaction to free up space include merging
keys, combining columns, and deleting tombstones. This process is managed by the class
