Vanilla 5 – New Features 2 – Vanilla Metadata Explorer

Dear Readers,

Another “reborn feature” : Metadata Explorer, available as an “early feature” in Vanilla 4, now fully functional, and integrated as a friendly user module. Using Metadata Explorer, any user can

• build his own request to extract data from the database (in fact, from the data he is allowed to access, as connection is through metadata only, and never direct access on database table)
• manipulate the dataset : add new column, create new calculated column, apply filter, aggregate information, drill between folders inside the metadata, etc …
• Save the dataview for later reuse (it will be available for edit/run from the portal), or save the dataset content as external file to be used by external packages like R or Weka (csv file, excel file, weka file)
• View the data using different integrated modules
  • graphs perspective : we provide the basic graphs view, a graph is worth hundred reports
  • generate a report on the data (switch to Vanilla Report keeping the current dataset),
  • generate a cube on the data, to be view using Vanilla Analysis viewer

Enjoy this series of screenshots …

Extraction Definition

DataSet Result

Graph Interface

Cube Definition

Cube Browser Interface

Report Interface

Now, using Vanilla Metadata Explorer, users can have access to their data, what ever the database is (relational database, Big Data database, NoSql database, etc …). Bi developer just need to design “best of the bread” Metadata, to allow users to access their business data.

Remember : Vanilla Metadata is an Open Source component, and provided as a full functional package (I mean : key features like security, multi join path definition, prebuilt queries, etc … all is available without any limitation), integrated as the Vanilla dataset provider with all the other Vanilla modules (Report, Dashboard, Cube, Kpi, Explorer, ETL …)

Have Fun !

Patrick

Vanilla Audit & Performance – Managing weight for requests (part 2)

Dear Readers,

Something usefull with Vanilla v4 is the possibility to manage weight for requests, and prevent the request to be issued against the database – based on predefined rules – if the weight exceed the predefined limit.

What is the interest ?

Using FreeWebReport, user are able to build complex report that contain crosstab elements and various grouping strategy. Some user need “assistance” when it comes to build complex reports, and sometimes, putting too many element in a report is not effecient. Purpose is to stop the request a user may raise before it goes to the database if Bi admin think the request is too heavy (simple example : too many tables in the sql statment).

How does it works ?

Each table in the Metadata document is set with a predefined weight. We all understand its not a database admin feature, but more a datawarehouse features : tables have different behaviour – fact tables, dimension tables, aggregate table – different number of records and different index strategy. – Using Enterprise Services, each group of the Vanilla Group Table is set with a global limit. – When a user run a request, the Vanilla platform calculate the sum of weight for all the tables included in the requests. If the sum is superior to the limit set to the group, the request is stopped and a message raised in the interface.

FreeMetadata – Setting Weight for Tables

Enterprise Services – Setting limit for Groups

This feature is very flexible, and allow you to set a strategy outside of the database landscape. This can be an explanation, as we will see in next post in another audit & control interface, why a request can’t run for a specific group

Have Fun !

Patrick

Vanilla v4 – Reminder about FreeMetadata

Dear Reader,

Its an evidence FreeMetadata is a must  in Open Source area ! Just ask JasperSoft … they dont have any document like that (how do you secure your dataset again : ok, I need to insert sql into a table … wow !) … Just ask Pentaho … ok, they have one … but who is using it … with which designer ? Again, I’m not talking about the Italian Slide xxx … because even the slide are crashing !!!!!!!!!

Its an evidence since 2008, we didn’t pushed enough FreeMetadata, because …. its running well ! Package is available since the first version of Vanilla and has proven to be robust, bug free, and completly Open Source (yes, its possible !)

Let’s be serious : FreeMetadata compete with major commercial products : integration with Dashboard, Olap Designer, Reporting (java & Web), Kpi Designer and even BiGateWay (Etl) … we provide security for groups, at row level, and at column level (column suppression and hide of content). Just walk through the different methods, with 2 groups : Default & Designer. Below, we show the result inside FreeMetadata – Designer Preview … in all the other designers, you just get the same resultset (yes, its ok with Birt … no, its not possible with IReport, simply because the IReport Eclipse development is just a marketing announcement !)

Group “Default” has a filter on “France”, and can’t see the “name” column

Filter definition on dateset

Very easy to definer simple or complex filter for a group … just imagine what you can use !

Column Management : we remove access for the group

Result

you dont have the column available for selection, and data are automitically filtered on ‘France’ 

Group “Designer” has no filter, and content of column “name” is dynamically hidden (we can call it “anonymisation”)

Column Management : we hide column content for the group

Result

We can see the column … but not its content !

What’s the interest to hide the content ? Just suppose you need to hire an external consultant to build report on sensitive data (yes, it happen in bank, financial department, police & military department, telco ….) and you dont want this consultant to read the value. This person need access to the colum to build the report, but you need to hide the column content …. just imagine how much you can save, because until now, you need to anonymise your database …. its a heavy job …

Have Fun !

Patrick

Vanilla & Birt

Dear Readers,

There is a key subject in Vanilla : if you like Birt & you know how to use Birt, you will like Vanilla & find it easy to use the various designers we provide (I’m talking about our bi Designers, such as FreeDashboard or our Birt Plugin for reporting).

In fact, Birt itself is already so nice & easy to use. But it lacks a Metadata abstraction level, a ready to deploy platform and is sometimes slow (due to its sql generation process)

Vanilla is provided with some crazy features, and we don’t push it as we should do (well, again, like other killing features in Vanilla). Vanilla server – that provide an easy way to deploy your Birt reports on Web – and Birt/FreeMetadata integration are one of those “killing features”, that allow you to have a choice between Actuate and Vanilla when it comes to Metadata & security integration.
You can draw a comparison grid (some companies did it), just don’t forget Vanilla FreeMetadata is available in Open Source, which means you can use it as part of our comunity edition to leverage your Birt platform (as opposed to Actuate commercial server)

Using Vanilla with Birt, you can take advantage of using a secured Metadata, but also enhance Birt report response time because Birt has response time limitation in its Open Source version (some of report’s Sql statment are not pushed to the database, which can lead to slow report production & heavy network trafic … ooooppsss !)

Using FreeMetadata with Birt is a simple workaround to Birt limitation (that force some companies to move to Actuate server), because FreeMeadata is the one that talk with the database, so it send a clean & optimisez Sql statment. In addition, using vanilla bi Portal, you will get crazy features such as dynamic column content, report history management, mail report delivery … just try it !

Vanilla/Birt is a nice result of Open Source product enhancement … and it was possible beacuse of the Eclipse foundation, that take care of having an open product that complies with standard. Coming back to last year, it was so sad to notice that another report package – Ireport – simply changed all its architecture and make it incompatible with previous version when moving from version 3 to version 3. Such a behaviour from the IReport development team is a key indication of the respect they have for the comunity working around the product to enhance it. It’s even more sad that this company has based its communication on the importance of projects coming from comunity … again a marketing position.

So, definitly, when it comes to Open Source Reporting, just keep Birt, and deploy your Birt report using a Vanilla server

Have Fun !
Patrick