what virtual cube look like, what pentaho virtual cube best are for, how to develop pentaho virtual cube, why virtual cube better that data blend, table-join based report, and also traditional pivot.
2. ABOUT THE AUTHOR
Currenty work as CTO at infinite-erp.co.id
Bring sized team consits of engineering
team to build and maintain high
performance information system.
You can reach me to
zaien@infinite-erp.co.id
Or visit my linked in
https://www.linkedin.com/in/zaien-aji-tr
ahutomo-83a63134/
CTO
I’m 100% BI tech big fan,
especially pentaho family. I love to
learn how to run pentaho at best.
With our love, expertise, and
background, we also encourage
you to gain this benefit, and join us
as a big Pentaho fans community.
BI Big Fan
I believe write it down cloud make
knowledge persists. I hope this
write not stop here, your turn to
make it deeper and better. If can
not, just share this to other, you
may find someone who can.
Just Write
3. What virtual cube look like? What animal is this?
What pentaho virtual cube best are for?
How software engineer bake it?
At the end we will arrive at conclusion: Why
pentaho virtual cube is better than data blend?
Better than table-join? Or even better than
traditonal pivot?
what
what for
how
why
01
02
03
04
ABOUT PENTAHO VIRTUAL CUBE
4. WHAT VIRTUAL CUBE LOOK LIKE?
Virtual cube my simmilar to traditional pivot. You will find this kind of analytics report in tradition BI. You will find some
dimensions and measurement, just like another dimensional report. Here our example show you sales and warehouse analysis.
You can see store dimension, and its several measurement. The first 3 dimensions talk abot sales performance, and next 3
dimension give you information related to warehouse. The last 3 dimensions give you insight that blend sales and warehouse
information such us profit that calculated by subtracting store sales to store cost.
5. WHAT VIRTUAL CUBE LOOK LIKE?
Behind the scene, we bake 2 analytics reports: not only sales analytics, but also warehouse analytics. Store dimension exists in
both reports, but promotions only exists in sales analytics. In measurement side, the first 4 measurmenet come from sales
analytics, and the rest from warehouse analytics. Thats why, when we drill down promotion dimension, warehouse analytics
have no data, then create a blank area. If this is JOIN BASED or DATA BLEND BASED report, then the 4th
row (promotion =
Best Saving) and consecutive records will not shown. Why? Because join based or data blend based report need keys to be
matched when they create report. THATS WHY VIRTUAL CUBE WIN!
This blank area is a proof
if this cube is not traditional cube
This is virtual cube
6. STRENGHTS
Performace matter! Since there is no
join, there are no overhead cost on
data processing and index scanning.
Make your report faster
WEAKNESSES
Virtual cube defined in mondrian
schema, make business user unable
define cubes them self.
OPPORTUNITIES
You can combine 2 or more analytics
report in single view. This
combination is UNLIMITED. You can
create any prespective, any angle,
on your analytics.
THREATS
Can you indentify any threats?
WHAT VIRTUAL BEST ARE FOR?
7. CAN YOU IMAGINE ANOTHER EXAMPLE?
ANOTHER
EXAMPLE
VIRTUAL
CUBE
1
2
3
4
In COA tree structure, you may define your
budget in summary level of chart account,
while your transaction happen in lowest level
of chart account. You will not be able to create
budget vs actual using data blend, since no
key (chart account key) matched
BUDGET VS ACTUAL
You may get sales performance declined not
because of demand get slow, but you may run
out of stock. You can simply create virtual cube
consists of sales and warehouse analytics, than
compare which SKU get run out of stock and
unachieved sales performance.
SALES PERFORMANCE VS STOCK LEVEL
You mat get sales performance declined not
because of demand get slow, but your
marketing dept not work well. You can simply
create virtual cube consists of sales and
expense analytics, than compare which SKU
get slow sales with slow campaign expense.
SALES PERFORMANCE VS PROMOTION EXPENSE
Any idea to exploit virtual cube?
YOUR TURN
8. HOW SOFTWARE ENGINEER BAKE VIRTUAL CUBE?
Under the hood, BI developer
define virtual cube based on
existing cube. They combine 2 or
more cube into single virtual cube.
The dimension may come from
common, or we so called shared
dimension. It’s measurement also
come from both side. As you can
see, this virtual cube can create
calculated member that come
both side, such as Profit per Unit
Shipped.
9. THERE ARE 2 SHARED DIMENSION
Can you spot another shared
dimension?
10. CONCUSION: WHY VIRTUAL CUBE BETTER?
PENTAHO VIRTUAL CUBE
OVER DATA BLEND, JOIN-TABLE
OR TRADITIONAL CUBE
2
Traditional report and virtual cube used in
similar way. End user may don’t know if they
work on several reports at the same time.
You may not event require additional
training to end user.
NO ADDITIONAL EXPERTISE
1
NO KEY JOIN REQUIRED
No key to be matched is matter. i.e you still
can create budget vs actual report if budget
is more general while actual are very detail.
4
It’s require 1 file only: mondrian schema!
Just define an xml tag “virtual cubel” then
start baking it’s capabilities through
existing cube.
EASY TO DEVELOP
3
PERFORMANCE
No key to be matched is matter. it’s just
several report rendered at the same time.
No key matching required, no index scan
performed. Report rendered faster.
11. READY TO BOOST PENTAHO BI
PERFORMANCE?
Head over to infinite-erp.co.id and find
out more about our background and
expertise.