HP vertica simple query optimization steps.


If you consider the basic steps to improve performance , please follow the below steps.

  1. Make sure tables are partitioned in a balanced way . Too many partitions are not good.The partision column needs to be set as NOT NULL. you can do it at any point using an ALTER TABLE command.
  2. Partition can be done by issuing an ALTER TABLE command
  3. PURGE_TABLE() for any purging of pending changes.
  4. Initially go for a projection update . You can select a complete initial projection without a specific script. This will make sure all the columns are included with super projections.
  5. Next step is to go and create a projection for the specific query using database designer tool. This should be done against incremental projections and deploy the projections.

 

Advertisements

Part 1 : Couchbase Server 2.0 – introduction


Couchbase server is a noSQL document based database for interactive web applications. It is ideal for web application which required low latency and high throuput. We can store data in a couchbase as keyvalue pair or  JSON document. Data doesn’t have to be following a fixed schema.

NoSQL database is a data storage mechanism which is highly optimized for retrieval and appending operation. It uses looser consistency model to achieve horizontal scaling and higher availability. NoSql dbms is very useful for managing huge quantity of data. Documents are addressed in the database using a unique key value representing a document.

Features of Couchbase

Flexible data model

JSON ( Javascript Object Notation) is a text based open standard designed for human readable data interchange

We can use JSON document to represents application objects and its relationships. We can have different data structure for the document based on the properties or type of that particular entity. Also we can modify the structure anytime without wanting to change the previously stored documents. It supports nested structure and field for representing relationship between items.

Easy scalability

Adding an additional node for handling more users is very easy so that we don’t need to change any application code. Also with one click of a button we can increase the cluster and distribute the load evenly. It supports automatic sharding of data and rebalancing. XDCR (Cross data centre replication) feature provide the facility to bring data from other data centre closer to the user.

Consistent High Performance

Massive concurrency and consistent high through put is what it is designed for. Server automatically distribute workload between nodes to maintain consistent high performance and also to avoid bottleneck at any server in the cluster.

Always online

It provides sub millisecond response time . XDCR and auto failover helps to ensure availability even for server failure or data centre shutdown.

 

ref: Couchbase documentation