Tuesday Feb 11, 2014

Filtered Charts

A customer question this week regarding filtering a chart. They have a report with a bunch of criteria with monetary values but, rather than show all of the criteria in a pie chart, they just want to show a few. For example:

 This ...
 rather than this

 There are a couple of ways to tackle this:

1. Filter the chart data in the chart definition. Using an XPATH expression you can filter out all of the criteria you do not want to see. Open the chart definition and update the definition. You will need to update the RowCount, RowLabels and DataValues attributes in the chart definition. Adding in the following XPATH expression:

    [DEPARTMENT_NAME='Accounting' or DEPARTMENT_NAME='Marketing' or DEPARTMENT_NAME='Executive']

so the DataValues value becomes:

    <DataValues><xsl:for-each-group select=".//G_1[DEPARTMENT_NAME='Accounting' or 
                                                     DEPARTMENT_NAME='Marketing' or 
                                                        DEPARTMENT_NAME='Executive']" ...

2. Create a variable in the template to hold just the values you want to chart.

    <?variable: filterDepts; /DATA_DS/LIST_G_1/G_1[DEPARTMENT_NAME='Accounting' or 
                                                     DEPARTMENT_NAME='Marketing' or 

Then update the chart definition with the variable for the same three attributes above, the RowCount, RowLabels and DataValues. For example:

    <DataValues><xsl:for-each-group select="$filterDepts" ...

These both work admirably, but they both require some manual updating of the chart definition which can get fiddly and a pain to maintain. I'm also just filtering for three departments, when you get up to 5 or 6 then the XPATH starts to become a pain to maintain. Option 2 alleviates this somewhat because you only need to define the filter once to create the filtered variable.
A better option may be ...

3. Force the effort down into the data layer. Create another query in the report that just pulls the data for the chart.

LIST_G2/G_2 holds the data for the chart. Then all you need do is create a vanilla chart on that particular section of the data.

Yes, there is some overhead to re-fetch the data but this is going to be about the same if not less than the extra processing required in the template with options 1 and 2. This has another advantage, you can parametrize the criteria for the user. You can create a parameter to allow the user to select, in my case, the department(s) they want to chart.

Its simple enough to create the multi-select parameter and modify the query to filter based on the values chosen by the user.

Sample report (including data model and layout template here) just un-archive into your catalog.
RTF Template plus sample data available here.

Monday Feb 10, 2014

Alternate Tray Printing

Since we introduced support for check printing PCL escape sequences in i.e. being able to set the micr font or change the print cartridge to the magnetic ink for that string. I have wanted to test out other PCL commands, particularly, changing print trays. Say you have letter headed paper or pre-printed or colored paper in tray 2 but only want to use it for the first page or specific or for a separator page, the rest can come out of plain ol Tray 1 with its copier paper.

I have had a couple of inquiries recently and so, I finally took some time to test out the theory. I should add here, that the dev team thought it would work but were not 100%. The feature was built for the check printing requirements alone so they could not support any other commands. I was hopeful thou!
In short, it works!

I can generate a document and print it with embedded PCL commands to change from Tray 1 (&l4H) to Tray 2 (&l1H ) - yep, makes no sense to me either. I got the codes from here, useful site with a host of other possibilities to test.

For the test, I just created a department-employee listing that broke the page when the department changed. Just inside the first grouping loop I included the PCL string to set Tray 1.

<pcl><control><esc/>&l4H </control> </pcl>

Note, this has to be in clear text, you can not use a formfield.
I then created a dummy insert page using a template and called it from just within the closing department group field (InsertPAGE field.) At the beginning of the dummy page I included the PCL string to get the paper from Tray 2:

<pcl><control><esc/>&l1H</control> </pcl>

When you run this to PDF you will see the PCL string. I played with this and hid it using a white font and it worked great, assuming you have white paper :)

When you set up the printer in the BIP admin console, you need to ensure you have picked the 'PDF to PCL Filter' for the printer.

If you dont want to have PCL enabled all the time, you can have multiple definitions for the same printer with/with out the PCL filter. Users just need to pick the appropriate printer instance. Using this filter ensures that those PCL strings will be preserved into the final PCL that gets sent to the printer.

Example files here. Official documentation on the PCL string here.

Happy Printing!

Friday Dec 20, 2013

FSG Reporting and BIP

This is a great overview of the Financial Statement Generator (FSG) engine from GL in EBS and how Publisher fits into the picture.Thanks to Helle Hellings on the Financials PM team.

Monday Nov 25, 2013

Conditional Borders

How can you conditionally turn cells borders on and off in Publishers RTF/XSLFO templates? With a little digging you'll find what appears to be the appropriate attributes to update in your template. You would logically come up with using the various border styling options:




Buuuut, that doesnt work. Updating them individually does not make a difference to the output. Not sure why and I will ask but for now here's the solution. Use the compound border formatter border-top|bottom|left|right. This takes the form ' border-bottom="0.5pt solid #000000". You set all three options at once rather than individually. In a BIP template you use:

<?attribute@incontext:border-bottom;'3.0pt solid #000000'?>
<?attribute@incontext:border-top;'3.0pt solid #000000'?>
<?attribute@incontext:border-left;'3.0pt solid #000000'?>
<?attribute@incontext:border-right;'3.0pt solid #000000'?>
<?end if?>

3pt borders is a little excessive but you get the idea. This approach can be used with the if@row option too to get the complete row borders to update. If your template will need to be run in left to right languages e.g. Arabic or Hebrew, then you will need to use start and end in place of left and right.

For the inquisitive reader, you're maybe wondering how, did this guy know that? And why the heck is this not in the user docs?
Other than my all knowing BIP guru status ;0) I hit the web for info on XSLFO cell border attributes and then the Template Builder for Word. Particularly the export option; I generated the XSLFO output from a test RTF template and took a look at the attributes. Then I started trying stuff out, Im a hacker and proud me!  For the users doc updates, I'll log a request for an update.

Thursday Nov 21, 2013

Desktop Testing XSL

Bit of a corner case this week but I wanted to park this as much for my reference as yours. Need to be able to test a pure XSL template against some sample data? Thats an XSL template that is going to generate HTML, Text or HTML. The Template Viewer app in the BI Publisher Desktop group does not offer that as an option. It does offer XSL-FO proccesing thou.

A few minutes digging around in the java libraries and I came up with a command line solution that is easy to set up and use.

1. Place your sample XML data and the XSL template in a directory
2. Open the lib directory where the TemplateViewer is installed. On my machine that is d:\Oracle\BIPDesktop\TemplateViewer\lib
3. Copy the xmlparserv2.jar file into the directory created in step 1.
4. Use the following command in a DOS/Shell window to process the XSL template against the XML data.

java -cp ./xmlparserv2.jar oracle.xml.parser.v2.oraxsl fileX.xml fileY.xsl > fileX.xls

The file generated will depend on your XSL. For an Excel output, you would instruct the process to generate fileX.xls in the same folder. You can then test the file with Excel, a browser or a text editor. Now you can test on the desktop until you get it right without the overhead of having to load it to the server each time.

To be completely clear, this approach is for pure XSL templates that are designed to generate text, html or xml. Its not for the XSLFO templates that might be used at runtime to generate PDF, PPT, etc. For those you should use the Template Viewer application, it supports the XSLFO templates but not the pure XSL templates.

If your template still falls into the pure XSL template category. This will be down to you using some BIP functionality in the templates. To get it to work you'll need to add in the Publisher libraries that contain the function e.g. xdo-core.jar, i18nAPI_v3.jar, etc to the classpath argument (-cp.)

So a new command including the required libraries might look like:

java -cp ./xmlparserv2.jar;./xdo-core.jar;./i18nAPI_v3.jar 
                            oracle.xml.parser.v2.oraxsl fileX.xml fileY.xsl > fileX.xls

 You will need to either move the libraries to the local directory, my assumption above or include the full path to them. More info here on setting the -cp attribute.

Wednesday Oct 23, 2013

BIP 11g Dynamic SQL

Back in the 10g release, if you wanted something beyond the standard query for your report extract; you needed to break out your favorite text editor. You gotta love 'vi' and hate emacs, am I right? And get to building a data template, they were/are lovely to write, such fun ... not! Its not fun writing them by hand but, you do get to do some cool stuff around the data extract including dynamic SQL. By that I mean the ability to add content dynamically to your your query at runtime.

With 11g, we spoiled you with a visual builder, no more vi or notepad sessions, a friendly drag and drop interface allowing you to build hierarchical data sets, calculated columns, summary columns, etc. You can still create the dynamic SQL statements, its not so well documented right now, in lieu of doc updates here's the skinny.

If you check out the 10g process to create dynamic sql in the docs. You need to create a data trigger function where you assign the dynamic sql to a global variable that's matched in your report SQL. In 11g, the process is really the same, BI Publisher just provides a bit more help to define what trigger code needs to be called. You still need to create the function and place it inside a package in the db.

Here's a simple plsql package with the 'beforedata' function trigger.


create or replace PACKAGE BIREPORTS AS 

 whereCols varchar2(2000);
 FUNCTION beforeReportTrig return boolean;




  FUNCTION beforeReportTrig return boolean AS 
      whereCols := ' and d.department_id = 100';
      RETURN true;
   END beforeReportTrig;


you'll notice the additional where clause (whereCols - declared as a public variable) is hard coded. I'll cover parameterizing that in my next post. If you can not wait, check the 10g docs for an example.

I have my package compiling successfully in the db. Now, onto the BIP data model definition.

1. Create a new data model and go ahead and create your query(s) as you would normally.

2. In the query dialog box, add in the variables you want replaced at runtime using an ampersand rather than a colon e.g. &whereCols.


select     d.DEPARTMENT_NAME,
 from    "OE"."EMPLOYEES" e,


Note that 'whereCols' matches the global variable name in our package. When you click OK to clear the dialog, you'll be asked for a default value for the variable, just use ' and 1=1' That leading space is important to keep the SQL valid ie required whitespace. This value will be used for the where clause if case its not set by the function code.

3. Now click on the Event Triggers tree node and create a new trigger of the type Before Data. Type in the default package name, in my example, 'BIREPORTS'. Then hit the update button to get BIP to fetch the valid functions.
In my case I get to see the following:

Select the BEFOREREPORTTRIG function (or your name) and shuttle it across.

4. Save your data model and now test it. For now, you can update the where clause via the plsql package.

Next time ... parametrizing the dynamic clause.

Thursday Oct 10, 2013

Mobile App Designer

Back in August a new Oracle mobile solution jumped out of the gate, the Mobile App Designer (MAD). I seem to have been on the road every week for the last, goodness knows how many weeks. I have finally found some time this week to get down and work with it. Its pretty cool and above all, its capable of providing a mobile platform independent reporting solution.

But you already have a mobile application! Yep, and I think they both sit quite comfortably together. The Oracle BI Mobile Application is available from the App Store for Apple users. Its a great app, build reports, dashboards and BIP reports for your browser based users and your Apple app users can take advantage of them immediately.

MAD takes the next step forward. Maybe you don't use or can not use Apple mobile devices? Maybe you need to build something more specific around a business area that provides users with a richer experience, beyond what Answers and Dashboards can offer. However, you do not want to have to rely of the tech folks to build the mobile application, thats just piling more work on them. You also want to be platform agnostic, you might have a mix of mobile platforms. MAD can help.

For those of you that have already used the Online Template layout editor with BI Publisher, you already know how to build a mobile application. The MAD interface is essentially the online template builder user interface, tweaked for a mobile destination ie a phone or tablet.

You build your data model as you would normally including the newer direct data model build on a subject area from OBIEE.

Then start to create the 'pages' of your application and the content to sit on those pages. All the normal stuff, tables, pivot tables, charts, images plus accordians, filters and repeating objects. On top of that is the ability to then extend the visual objects that are available to users. Maps (google or oracle), D3 visuals, gantt charts, org charts, if you can either write the code or leverage an existing javascript library, MAD has the extension framework to support it.

You can build and test in a browser and then deploy to your own BI App Store. Users, on their mobile devices, can then subscribe to an application. They can open and interact with your app using their phone or tablet's interactive features just as they would with a native application.  As you update your app and add new features the changes will be picked up the next time your users open the application.

Interested? Want to know more? The Oracle MAD home page has a ton of content including tutorials, etc. We are planning to dig into MAD in forthcoming posts. The geek in me wanted to be able to build plugins using the D3 and other visuals. I have been working with Leslie on some of the documentation and we'll be sharing some of that 'plugin' doc and how tos in the coming weeks.

Monday Jul 15, 2013

Minning and Maxing in Pivots

A tricksy question from a hobbiteses this past week or so. How can I use minimum or maximum in an RTF template pivot table?

Using the pivot table dialog box, you get sum or count. So, how to get a min or max? You need to understand the pivot structure a bit to understand how to get the min|max. I wrote about the pivot table format a few years back here.

 Its the C field that holds the calculation as the last parameter.

<?crosstab:c8949;"//G_1";"DEPARTMENT_NAME{,o=a,t=t}";"HIRE_YEAR{,o=a,t=t}";"JOB_ID";"sum" ?>

I was not sure if we could simply swap out the sum|count function for our min, max functions. But, Im a hacker at heart, so I gave it a whirl. It worked, I used the BIP min and max functions:


They both work nicely!

So, the C field would look like:

<?crosstab:c8949;"//G_1";"DEPARTMENT_NAME{,o=a,t=t}";"HIRE_YEAR{,o=a,t=t}";"JOB_ID";"xdoxslt:maximum" ?>

If you do not need the default totals (that use the functions you define.) You can just delete them from the table.

Sample template and data here.

Now, the average values need cracking!

Thursday May 23, 2013

BI Publisher 11g Training - Jul 1-3

For those of you still sitting on the 10g fence ... or if you're new to Publisher 11g, take advantage of this educational opportunity:

Oracle BI Publisher 11g R1: Fundamentals

Learn To:

    Create data models by using the Data Model Editor.
    Create BI Publisher reports based on data models.
    Create report layouts by using the Layout Editor (online).
    Create reports based on OBI EE data sources.
    Publish the reports on OBI EE Dashboards.
    Schedule reports and burst these reports.

Date: 01-JUL-13
Duration: 3 days
Location: Online

Click here for more details and to enroll

Friday May 10, 2013

Building on Subject Areas

The new release of BI Publisher has a very nice new feature for those of you wanting to build reports on top of the BI Server data model. In previous releases you would need to either write the logical sql yourself or build an Answer request and copy the SQL from the Advanced tab and paste it into the BIP data modeler.

With the new release comes the ability to create reports without the need for a data model at all. You have the option when creating a new report to use a subject area directly.

 Once you have selected the subject area you are interested in you can decide on whether to continue into the wizard to help you build the layout. Or to strike out on your own and build the layout yourself.

If you go for the latter and load up the layout editor, you get to see all of the data items you would see in the Answers builder in the data tree. Its then a case of dragging and dropping the columns into the layout, just as you would normally with a sample data source.

Once you are back to the report editor, the final step is to add some parameters. 

This is a little different to a conventional BIP report. There is no data model definition per se i.e the logical SQL is not stored but rather, the columns you added to the layout and the subject area(s) you pulled them from. Yes' you can go across subject areas, but you need to know if its going to make sense or even work before you add more. You add more subject areas click on the subject area name where the data model name normally resides. You'll then get a shuttle dialog that lets you add more subject areas. You can then add columns in the layout builder.
Getting back to the parameters, on the report editor page, click the Parameters link (top right.) This will open the parameters dialog.

You can add parameters and set how they will be displayed; whether folks can select all; do they see check boxes, a drop box or text box; whether other parameters should be limited by the choice made for this box. Everything you get with a regular BIP parameter.

Finally, the report rendered with the parameters.

If you have a need to build a more highly formatted report on the BI Server data then this is definitely the way to go. This approach really does open up BIP reporting to business users. No need to write SQL, just pick the columns you want and format them in a simple to use interface.

Before you ask, you can not build report layouts in MSWord or Excel for this type of data source, not yet anyhoo :0)


Follow bipublisher on Twitter Find Us on Facebook BI Publisher Youtube ChannelDiscussion Forum

Join our BI Publisher community to get the most and keep updated with the latest news, How-to, Solutions! Share your feedback and let us hear your voice @bipublisher on Twitter, on our official Facebook page, and Youtube!


« April 2014