UnitedOlap, Designer Explorer

Dear Readers,

First of all, I would like to thanks all the persons who contacted us to be part of the UnitedOlap test Team. It’s really a nice message sent to the UnitedOlap dev team to have experienced persons behind us to finalise some features and run tests on various databases.

I took time to answer to most of the mails (for those with clear mail address), to indicate that we forecast to add new testers with period 2 of the global test, which should start second half of july, once we will have collected & solved most of the identified bugs.

During the test period, development of new features such as new integrated explorer to browse more objects while developing the schema. I will go short on the competition, because there is nothing in Open Source (Designer + EXplorer of Cubes), leading to painfull cycle of development/tests.

So, thans to Eclipse PlugIn technology, we have added 2 explorers in our FreeAnalysis Schema Designer, one basic, to test direct Mdx statment, and another one with more interactivity, which purpose is to take place of FreeAnalysis Java, our java cube package.

Interface to select dimensions & measures

2 interestings features : you can select your profile/group to browse the dataset and you can limit the number of rows used from the fact tables (during the develpment/test, to reduce time to build the grid)

Results in a grid on large dataset

Revival of the Java package …. !!!

We finally make it … still under development, but already usable, promising and soon flashy !

… dont miss a chance to be part of the next generation of Open Source Olap Engine, join us and be involved in the Roadmap !

Have Fun !

Patrick

Vanilla Viewer

Dear Readers,

There is a nice piece of software available in Vanilla 3.4 : Vanilla Viewer, and this post will echo my previous post, 25th avril, on this subject. Basically, Vanilla viewer takes place of Birt Viewer, but provides more : for users, a common visual interface that support Birt & Ireport reports (and we can imagine running also other report format …). This means : no matter what is the developer interface, user will access a standard unified interface to run it.

Vanilla Viewer interface

Second point, we provide more : OpenOffice “save as …” features, that allow users to save the report in Oo format. Also, Vanilla Viewer support standard features such as cascading prompts and multi selection, along with the possibility to save the report in our document workspace, for indexation and reusability (something which is missing i Birt Viewer as you can only save your report in local, and not on your portal’s workspace.)

Interface with save option (internal to Vanilla document system)

Vanilla Viewer – cascading selection

All in one, the direction for Vanilla Viewer is to provide more support for FreeAnalysis view, FreeDashboard document, statistical values for BiGateway transformation and BiWorkflow process …

Vanilla Viewer – Pdf Export

Beeing developped as a component, Vanilla Viewer is easilly embedable in any external portal, as a portlet, to provide external access to Vanilla Bi documents,

Have Fun !

Patrick

Update for Vanilla Apps

Dear Readers,

Early this week, we make public new “ready to install & deploy” packages for OpenBravo (Erp) and VTiger (Crm) platforms, that complement the existing package Vanilla4Cockatil (Open Source Erp for Universities) and Vanilla4Mantis (Support platform).

Vanilla4OpenBravo is a collection of reports, cubes and dashboards built on top of the OpenBravo database structure, that provide immediate access to information stored in the database. As all the Bi documents are built using a FreeMetadata document, it’s very easy to customise to your security policy, and even use the metadata as a virtual datasource when extracting data with BiGateWay Etl.

Vanilla4VTiger is exactly the same approach and spirit, but on top on Vtiger database structure.

More information can be found on Bpm-conseil Web Site, and packages are available on our Forge.

Have Fun !

Patrick

UnitedOlap and Manual Dimensions

Dear Readers,

Another crazy feature is available in UnitedOlap, using our UnitedOlap Designer : you can easilly create a dimension level by your own (we call it MANUAL DIMENSION) and assign your child member to “virtual father node”. It’s so usefull when you miss a new level to group your data.

Just  let me give you a simple example. Suppose you have a list of countries but not the “continent” information in your database. Which means by default you can’t view the data per continent without any heavy operation such as adding a new table in your database, or a new column in your dimension table, along with building an interface to enter the codification.

With UnitedOlap, simply create MANUALLY a new dimension level (“continent” in our example) and assign each country to its continent. In 10 minutes, you can view your data using a “continent” dimension, without any new tables in your database (or SpreadSheet, / csv datafile) … all is under control by the Olap Designer …

… soon available in preview, this summer, along with documentation (Pdf, Video) …

Have Fun !

Patrick

Talking about BigData

Dear Reader,

It’s been a few months that we can read a lot about no-sql database, columns database, … and most posts dont make the difference between the storage strategy and the language used to exchange with the database engine.

Talking about Business Intelligence, we tried to find an interesting subject to test a BigDatabase engine (Casandra, part of the Hadoop Project) with the Vanilla platform (which is, by nature, sql-oriented)

Purpose is to understand what are the limitation of no-sql language when using Reporting or Etl package, how we can connect Vanilla to no-sql database and what are the performance when compare with a sql-database (PostGreSql), so that we can help our customers and partners when talking about “fashion” BigData projects.

One of the most interesting feature of Vanilla 3.4 is the Maps module, embeded into Dashboard, Report, Kpi and Cubes. So, it was a natural situation to test Vanilla and Cassandra inside a Geolocalisation Projects.

Honest , beginning of project was not trivial, as we had to turn our mind from “sql requests” to adopt another strategy (java class to manipulate data), and we even maintain a PostGreSql database structure in parallel to validate storage content and requests results. Finally, first result is impressive, as Cassandra allow us to mix objects (such as images, maps) and data (such as kpi, location properties) in a common space and provide a single entry for “data” manipulation. Still prototyping the projects, it was very pleasant to be able to add new objects types without running heavy database structure modifications.

In addition, performances are great AND stable : no matter if we have 1 million of data or 20 millions inside Cassandra, result of geo-requests are stable (some milliseconds to get complex calculation that return a set of geodata properties).

We will soon make our development available for the mass, just keep in touch,

Have Fun !

Patrick