<!--
Licensed to the Apache Software Foundation (ASF) under one
or more contributor license agreements.  See the NOTICE file
distributed with this work for additional information
regarding copyright ownership.  The ASF licenses this file
to you under the Apache License, Version 2.0 (the
"License"); you may not use this file except in compliance
with the License.  You may obtain a copy of the License at

  http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing,
software distributed under the License is distributed on an
"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
KIND, either express or implied.  See the License for the
specific language governing permissions and limitations
under the License.
-->

# Language change : from Thrift to CQL

The data model changes when using *CQL* (Cassandra Query Language).
*CQL* is providing an abstraction of the low-level data stored in *Cassandra*, in
opposition to *Thrift* that aims to expose the low-level data structure directly.
[But note that this changes with Cassandra 3’s new storage engine.](http://www.datastax.com/2015/12/storage-engine-30)

*Thrift* exposes *Keyspaces*, and these *Keyspaces* contain *Column Families*. A
*ColumnFamily* contains *Rows* in which each *Row* has a list of an arbitrary number
of column-values. With *CQL*, the data is **tabular**, *ColumnFamily* gets viewed
as a *Table*, the **Table Rows** get a **fixed and finite number of named columns**.
*Thrift*’s columns inside the *Rows* get distributed in a tabular way through the
*Table Rows*. See the following figure:

```ditaa
                                                     Thrift
             /-                                                                                          -\
             |                                                                                            |
             |  /------------\              /---------------+---------------+---------------+---------+   |
             |  |    cRED    |              |cFA0    1      |        2      |        3      |         |   |
             |  |     1      | ---------->  +---------------+---------------+---------------+ ...     |   +--> One Thrift
             |  |            |              |c1AB   'a'     |       'b'     |       'c'     |         |   |    ROW
             |  \------------/              \---------------+---------------+---------------+---------+   |
             |                                                                                            |
One Thrift   |                                                                                           -/
COLUMNFAMILY |          
             |  
             |  /------------\              /---------------+---------------+---------+
             |  |            |              |        1      |        2      |         |
             |  |     2      | ---------->  +---------------+---------------+ ...     |
             |  |            |              |       'a'     |       'b'     |         |
             |  \------------/              \---------------+---------------+---------+
             |  
             \- 


                      -----------------------------------------------------------------------


                                                     CQL

             /-                              
             |                               
             |  /--------------------+---------------------------------+-----------------------------\
             |  |       key          |              column1            |            value            |  
             |  +--------------------+---------------------------------+-----------------------------+
             |  | cRED   1           |  cFA0          1                |   c1AB      'a'             |
             |  +--------------------+---------------------------------+-----------------------------+ -\
             |  | cRED   1           |                2                |             'b'             |  +--> One CQL
   One CQL   |  +--------------------+---------------------------------+-----------------------------+ -/    ROW
   TABLE     |  | cRED   1           |                3                |             'c'             |
             |  +--------------------+---------------------------------+-----------------------------+
             |  | cRED  ...          |               ...               |             ...             |
             |  +--------------------+---------------------------------+-----------------------------+
             |  |        2           |                1                |             'a'             |
             |  +--------------------+---------------------------------+-----------------------------+
             |  |        2           |                2                |             'b'             |
             |  +--------------------+---------------------------------+-----------------------------+
             |  |       ...          |               ...               |             ...             |
             |  +--------------------+---------------------------------+-----------------------------+
             \-
```

Some of the columns of a *CQL Table* have a special role that is specifically
related to the *Cassandra* architecture. Indeed, the *Row key* of the *Thrift Row*,
becomes the *Partition Key* in the *CQL Table*, and can be composed of 1 or multiple
*CQL columns* (the key column in Figure 1). The  *“Column”* part of the Column-value
component in a *Thrift Row*, becomes the *Clustering Column* in *CQL*, and can
also be composed of multiple columns (in the figure, column1 is the only column
composing the *Clustering Column*, but there can be others if the Thrift’s ColumnComparator
is a CompositeType).

Here is the basic architectural concept of *CQL*, a detailed explanation and *CQL*
examples can be found in this article : [http://www.planetcassandra.org/making-the-change-from-thrift-to-cql/](http://www.planetcassandra.org/making-the-change-from-thrift-to-cql/).
Understanding the *CQL* abstraction plays a key role in developing performing
and scaling applications.
