Infokus

 You can download this tutorial as a PDF by clicking this picture!

This is article is written based on below queries before I explore it.

1) Is it mandatory to have Star Schema based warehouse to create Cubes ?
2) In oracle if you have multiple users(lets say multiple schemes).. Is it possible t design a simple CUBE dimension from schema A of a table and measure from schema B of another table ?

The answer is , NO & Yes.

All the organizations may not maintain complete data ware housing solutions and run their transactional database using normalized OLTP models.

It is not mandatory to have a star schema based ware house to create OLAP cube to analyze your data but you can also use OLTP databases.


The new things I found while demonstrating a simple cube are..

1) Connecting to oracle database.
2) Fetching all the tables of all schemas.
3) Synonyms of a particular schema

 You will find a lot of time taking to load the database into PSW tool once you connected to a schema.. Using FILTER_SCHEMA_LIST at Options section of the database dialogue will take the control to load only specified schema tables.

By default the PSW engine assumes to load all the tables in all the schemes.


A scenario of Dimensions and measures.

Suppose in a case, if you have a measure coming from table defined in SchemaA, measure coming from a table defined in SchemaB... in that case while giving tables for fact and dimension tables, the synonyms will not appear in drop down boxes... 

You need to hard code the tables names ... Of course your dimensions & measures will be in X with red color, it will not affect your definition of hard coded table names.

Save the schema and publish it to the server... have tested it with Saiku plug-in ad-hoc editor and worked nicely.


Thank you.

References :
1) http://jira.pentaho.com/browse/PSW-135

2) http://forums.pentaho.com/showthread.php?162540-OLTP-tables-VS-RDBMS-tables-Designing-OLAP-cubes-Peformance-factors&p=374714#post374714

3) http://forums.pentaho.com/showthread.php?157997-Schema-Workbench-freeze-after-establishing-a-connection-in-db








 

Hi Guys,

I love experimenting/Playing with Pentaho C-Tools where it took me to explore D3 bar chart with SQL query result set...

Not a big deal in implementing it, but want to share some thing with community folks.

What will you learn in this post ?
You will get an idea on how to implement D3 bar chart.
1) Passing query output to D3 chart code.



Limitations on the Example :
1) Query and Chart is not parametrized yet.
2) Y-Axis labels are not dynamic yet.
3) Tool tip is not implemented.
4) Not tested the bootstrap support.
5)  and more.....

Readers are encouraged add your support to improve the sample of this in comment box...


Software setup :
a) C-Tools(CDE,CDF,CDA,CGG) - Version14.06.18 stable with bootstrap support.
b) Database server for a sample query - In this example postgres - foodmart database.
3) Install D3 components plug-in from market Place

 Example 1 : 

 Step 1 : Layout : Create your layout .... Row---column--Html... Span 24 for this example
 Step 2 : Data Source : 
Driver : org.postgresql.Driver
URL : jdbc:postgresql://localhost:5432/foodmart
UserName/Password : postgres/postgres
Query : query1
select distinct first_name as letter,salary as frequency from employee where salary!=20 limit 20
Step 3: Components
D3 Components - D3 component -
In properties for Custom Chart Script write below code.
Datasoruce : query1


function f(dataset){
   

    
    var data = this.cdaResultToD3Array(dataset);
   
    
    var margin = {top: 20, right: 20, bottom: 30, left: 40},
        width = this.getWidth() - margin.left - margin.right,
        height = this.getHeight() - margin.top - margin.bottom;
   
    var formatPercent = d3.format(".0%");
   
    var x = d3.scale.ordinal()
        .rangeRoundBands([0, width], .1, 1);
   
    var y = d3.scale.linear()
        .range([height, 0]);
   
    var xAxis = d3.svg.axis()
        .scale(x)
        .orient("bottom");
   
    var yAxis = d3.svg.axis()
        .scale(y)
        .orient("left")
        .tickFormat(formatPercent);
  
  
    var svg = d3.select("#"+this.htmlObject).append("svg")
        .attr("width", width + margin.left + margin.right)
        .attr("height", height + margin.top + margin.bottom)
      .append("g")
        .attr("transform", "translate(" + margin.left + "," + margin.top + ")");
       
       
 
   
    // Commenting this out, we have data already
    // d3.tsv("/pentaho/api/repos/d3ComponentLibrary/static/custom/data/data.tsv", function(error, data) {
   
      data.forEach(
                    function(d) {
                    d.frequency = +d.frequency;
                    }
      );
   
      x.domain(data.map(function(d) { return d.letter; }));
      y.domain([0, d3.max(data, function(d) { return d.frequency; })]);
   
      svg.append("g")
          .attr("class", "x axis")
          .attr("transform", "translate(0," + height + ")")
          .call(xAxis);
   
      svg.append("g")
          .attr("class", "y axis")
          .call(yAxis)
        .append("text")
          .attr("transform", "rotate(-90)")
          .attr("y", 6)
          .attr("dy", ".71em")
          .style("text-anchor", "end")
          .text("Frequency");
   
      svg.selectAll(".bar")
          .data(data)
        .enter().append("rect")
          .attr("class", "bar")
          .attr("x", function(d) { return x(d.letter); })
          .attr("width", x.rangeBand())
          .attr("y", function(d) { return y(d.frequency); })
          .attr("height", function(d) { return height - y(d.frequency); });
   
      d3.select("input").on("change", change);
   
      var sortTimeout = setTimeout(function() {
        d3.select("input").property("checked", true).each(change);
      }, 2000);
   
      function change() {
        clearTimeout(sortTimeout);
   
        // Copy-on-write since tweens are evaluated after a delay.
        var x0 = x.domain(data.sort(this.checked
            ? function(a, b) { return b.frequency - a.frequency; }
            : function(a, b) { return d3.ascending(a.letter, b.letter); })
            .map(function(d) { return d.letter; }))
            .copy();
   
        var transition = svg.transition().duration(750),
            delay = function(d, i) { return i * 50; };
   
        transition.selectAll(".bar")
            .delay(delay)
            .attr("x", function(d) { return x0(d.letter); }
           

            );
           
        transition.select(".x.axis")
            .call(xAxis)
          .selectAll("g")
            .delay(delay);
      }
    // });
   
}
NOTE :
1)  As the code is a replica of build in Example in Pentaho D3 plug-in , the code depends on 2 column result set of data.....

Step 4: Save dashboard and test the output
Image 1 : With sort enabled - For this you need to write below HTML code for the column of this chart in layout section .

<label><input type="checkbox"> Sort values</label>  


Image 2 : With sort check box disable

Example 2:

Why to wait ? Lets have a look at this static bar chart example in CDE( check this : http://jsfiddle.net/enigmarm/3HL4a/13/ )

Image 2 : Descending
Default:

To get the above static charts on CDE output... lets do this..

Hmmmm...!!! the same  code I used from http://jsfiddle.net/enigmarm/3HL4a/13/ in CDE as an experiment..

Add HTML code at Row-Column-HTML
Add Css code at : Code Sinppet Css
Add D3 chart script code at : Custom chart script for D3 component.
Save your dashboard and preview it..


Are you eager to check it in CDE ? Download the examples here 

Click Me folk...!!! :-) 


References : 
1) http://bl.ocks.org/Caged/6476579
2) http://bl.ocks.org/mbostock/raw/3885705/
3) http://bl.ocks.org/mbostock/3885304
4) http://bl.ocks.org/mbostock/raw/3885304/
5) http://jsfiddle.net/gregfedorov/Qh9X5/9/
6) http://jsfiddle.net/uedatakuya/tXPEV/light/
7) http://jsfiddle.net/weXNd/6/