Monday, September 2, 2013

RTP software symposium spring data

Talk by Craig Walls
Redis key value store, Riak,...
Spring data umblla project 3 years ago
JPA, Spring android, MongoDB, CouchDB,Neo4j, HBase, Cassandra, jdbc extensions, Hadoop, Gem Fire data grid, amazon s3 blob, rest exporter,
a lot data harvested from somewhere based on that make decisions, i.e election and campaign

Traditional database not capable holding this much data,  distributed file system, send workers at it like mapreduce, with Hadoop , gemfire is like a key value store, redis key value memory you could use it as session,  jms heavy you can do messaging over Redis, document store highly denormalized  and no schema, data not related, storing big chunks of data, resume data is an example, json like , graph databases is great neo4j normalize it into nodes and have relationships easy to understand , column stores key value within a key value, not really worked on Cassandra anymore, focus on hbase, blob s3 not spending time on it

Template oriented data access, auto magic repositories, annotating objects so u can store it

originally a fan of castor , then hibernate came, saved objects, JPA formalized hibernate, avoid queries, he didn't like it, struggling at mappings, complex data gets hard, for example do I make it eager or lazy?, don't like , rather write queries than deal with this, no JPATemplate, suggest not using HibernateTemplate, and JPATemplate, raw form hibernate better, jPA repositories is what is useful,

orderRepsitory  interface extends JPARepository <Order, Long>

No implementation and his test case passed


<JPA:repositories base-package = "">

Find respository method naming

And, Or, Betweeb, LessThan, GreaterThan, IsNull, IsNotNull, ! NotNull, Like, NoLike, OrderBy, Not , In, NotIn,


Annotate as Document

Mongo says Ids need to be String

Infinatest  plugin,

<mongo:repositories base-packages = "">


NodeEntity annotation
Lucene under covers

No comments:

Post a Comment