Last release: DEX 4.8
Notice that this release has an important change in the data format. Before switching to 4.8 you should backup your current database and then restore it with the new version.
For more information about the backup & restore functionalities, take a look at the Maintenance & Monitoring chapter from the User Manual.
Improved performance
- Faster shortest path algorithm: We have improved the BFS Shortest Path algorithm to get quicker results.
- Bitmaps management: Improvements in the internal structure to boost the performance.
- Better concurrency: Better scaling with concurrent sessions.
Older versions
DEX 4.7
DEX 4.7 into HA
- Horizontal scaling: Increases the throughput (queries/second) as many times as the number of DEXHA nodes in the HA system.
- Availability: The replicated system available in DEXHA mode will help you assuring a high service level agreement.
New traversal algorithm
- Context alogrithm: Context is a new traversal algorithm which retrieves all the visited nodes in an Object instead of using the iterator. More information about the use of the new algorithm in each reference manual: Java / .Net / C++
Scripting improvements
- Define attributes: Attributes can be created with independency of the node or edge definition. Attributes can also be defined as global.
- Global attributes:scripting commands (drop, index, etc..) now available for attributes as well. More information about the use of the scripts in each reference manual: Java / .Net / C++
Previous definition for attributes was like the following:
create node PEOPLE ( ID int, NAME string basic )
New definition for the same example would now be:
create node PEOPLE
create attribute PEOPLE.ID int
create attribute PEOPLE.NAME string basic
Improvements in String statistics
- Maxium length, minimum length, mean length
DEX 4.6
DEX 4.6 makes DEX aCiD.
- Durability: Changes will persist thanks to the complete Recovery Manager. The recovery manager keeps your DEX databases automatically backed up all the time, and provides recovery tools in case you may need to delete unfinished Tx.
- Consistency: After each Tx the GBD is consistent, guaranteed with the operations order.
- Atomicity: Better autocommitted transactions (not rollback yet).
- Isolation: Simple isolated Tx with S/X blocking.
DEX 4.5
DEX 4.5 includes the following new features:
- Graph Algorithm package for all APIs
Graph Algorithm package includes the following algorithms:- Traversals algorithms: To traverse the graph using DFS or BFS techniques. Your choice!
- Find shortest path algorithms: Find the shortest way to between two nodes, using BFS or Dijkstra techniques, whatever fits your code best!
- Connected components algorithms: Find Strongy or Weakly connected components.
DEX 4.4
DEX 4.4 includes the following new features:
- C++ API
-
Fully C++ API.
DEX is the only graph database giving you the highest performance in C++. - Backup and restore funcitionalities
-
Keep your graph database save and do not worry to any change;
the logical backup and restore makes it possible to change to any different platform or DEX version.
Even more, the restoration is done in less than the half of time of loading your database. - New Schema dump scripting
- Dump your schema in an script format which is ready to be imported.
- Improved cache statistics
-
More usable statistics, to know everthing happening in your cache.
DEX 4.3
DEX 4.3 is distributed in 3 different APIs:
- JAVA: Java API for DEX. We have remodelled the Java API for a better structured future of DEX.
- New feature - DEX Java 4.3 includes loaders and exporters for edges and nodes
Allows creating a graph and manipulate its schema using a script. Load nodes and edges direclty from a csv file
Take into account that currently this API is not offering graph algorithms, they will be up in the next release.
- New feature - DEX.NET 4.3 includes loaders and exporters for edges and nodes
Allows creating a graph and manipulate its schema using a script. Load nodes and edges direclty from a csv file
Take into account that currently this API is not offering graph algorithms, they will be up in the next release.
DEX 4.2
DEX 4.2 is distributed in 3 different APIs:
- JAVA: The new Java API for DEX. We have remodelled the Java API for a better structured future of DEX.
Take into account that currently this API is neither offering graph algorithms nor script loaders, they will be up in the next release. - .NET: The brand new API makes possible to Microsoft .NET programmers to use the high performance of DEX graph database.
- JDEX: For a complete compatibility with applications from previous versions of DEX, we are keeping the Jdex API, which also includes improved log errors messages.
DEX 4.1
In addition to the features available at DEX 4.0, version 4.1 adds:
- Compatibility with Windows 64 bits
- Compatibility with MAC OS Snow Leopard
- Exporting DEX graphs to standard GraphML
- More efficient use of space
- Multi-level log system
- Allow creating iterators to navigate through data skipping some objects
- Other minor improvements for objects and types
DEX 4.0
In addition to the features available at DEX 3.0, v 4.0 adds:
- Achieve even better performance results for very large volumes of data: up to 5 billions of objects.
- Variety of licenses depending on the size of the database.
Licenses up to 1 million objects, up to 100 million objects, up to 1 billion, or more than 1 billion.
Contact us and we will see which one will suits you better. - Database configuration is made by default there's no need to manually configure any parameter.
System manages memory configuration as a function of the workload. - Cross-platform compatibility. Moving your database from one platform to another is just as easy as copying a file.
Compatibility with platforms with the same endianness even between 32 and 64 bit platforms. - Statistical analysis of information from attributes.
Up to 11 statistics available for attributes: check how many attributes have the same value or other typical statistical information such as mean, variance, median.... - Partial ACID support for transactions.
- Read-only database mode to avoid unexpected updates.
- Explode a collection of nodes with just one operation. Neighbors operation also available for a collection of nodes.
- Create edges by value, you do not need to get the identifiers of the tail and head but create edges by attributes value.
- Post-indexing of attributes: create first your attribute and index it later.
- More edges and values for attributes thanks to internal structures improvements.
DEX 3.0
- Graph algorithms such as shortest path, connectivity and traversal.
Read more: http://sparsity-technologies.com/blog/?p=97 - Regular expressions for queries.
Read more: http://sparsity-technologies.com/blog/?p=136 - Indexed or not indexed attributes.
Read more: http://sparsity-technologies.com/blog/?p=162 - Materialized neighbors.
Read more: http://sparsity-technologies.com/blog/?p=155 - New large char objects type created.
Read more: http://sparsity-technologies.com/blog/?p=166 - Define export for nodes and edges types (also for virtual edges types).
There's no need to define an export for each instance in the database, instead it can be defined once for all the objects of the same type.
In addition exports can be created both at instance level and at type level. - "Between" operation made available for querying.
- Native library included in the jar library and automatically loaded from there.
No need to take into consideration native library when exucting a program. Just execute DEX as any other regular java library. - Optimization of queries.
- Integrity restriction to avoid duplicates when creating an index for an attribute.
- Stream append operation. This operation is complementary to creating a new type for large char objects.
Append operation allows adding information to those objects appended from the end. - Objects sample method to obtain a random sample from a certain group of objects to work with.