The OLAP Printer sounds very interesting, and you are probably right
that we should look into OLAP4J.
Best,
Knut
---------- Forwarded message ----------
From: Jiri Dvorak <Jiri....@t4bi.com>
You are correct to say that we need to coordinate all these components
and technologies closely, so that we'll end up with a consistent and
reusable framework. We are currently in the process of wrapping up
the 1st phase of OH2 component development, and we should have both
the code and the documentation ready in 2-3 weeks; I should be able to
send you a preliminary documentation (TA document) draft possibly even
next week.
The current WHOSIS code I gave you in Geneva has 2 limitations:
- JSON renderer has to be added (by reusing the code for the XML.V1
and/or Excel/CSV renderers)
- The XML/A interpreter has been tested only for 3 dimensions
(Country, Year, Indicator)
If you can live with 3 dimensions in the cube, then adding a JSON
renderer should be only a moderate amount of work.
Just for completeness, if you are planning on building most of the
"final" UI yourselves on the project, and you simply need to send an
MDX query to Mondrian and get a bunch of result rows back, the
standard library used by Pentaho, JPivot and others is OLAP4J
(http://www.olap4j.org/), where the API has been modelled after JDBC
queries (only, you specify MDX query strings instead of SQL ones) -
so, any Java programmer who worked with JDBC should find it easy and
intuitive to use. However, this API gives you only "raw data",
nothing else.
On our end, for the OH2 project, we will have several tools and APIs
ready. The fully interactive OLAP client is implemented in Flex, but
we also have a separate RESTful server (called "OLAP Printer"), which
can accept MDX queries (*NOT* limited to 3 dimensions or specific
cubes), and return data grids in several formats (currently PDF and
Excel, planning to add easy-to-parse HTML shortly). The results
produced by "OLAP Printer" will be compatible with the interactive
OLAP client, including e.g. handling of aggregates and WHO "metadata"
(footnotes) or links to charts. I have added Dusan Strnal, the main
"OLAP Printer" developer, to the CC list of this thread, if you have
more questions - or maybe you can read the OH2 Technical Architecture
document first (next week, or so).
Take care
- Jiri
----------------------------------------------------------------
Jiri Dvorak, T4Bi (www.t4bi.com), cell +1.604.220.0529
#1010-1177 West Hastings Street, Vancouver, BC V6E2K3, Canada
-----Original Message-----
From: Knut Staring [mailto:knu...@gmail.com]
Sent: Wednesday, January 07, 2009 5:27 AM
To: jiri....@t4bi.com; Paul Pimstone; Veltsos, Philippe; Lemarchand, Johan
Cc: Lars Helge Øverland; Ola Hodne Titlestad; Jan Henrik Øverland
Subject: Happy New Year
Dear Paul and Jiri,
Happy new year - I trust 2009 will be good for you and for our
collaboration! I hope you found some time to relax over the holidays!
Having been busy with other things (trying to wrap up my phd in a
month or two), I would like to follow up on the OLAP side of things.
We have done some work on linking MapFish to the DHIS using web
services, but I would very much like to try and make use of your new
OLAP code, as we discussed in Geneva in October.
The code we have is what you gave us then, do you recommend working
with that, or are there some new developments you would like to share
with us? If we want to hook it up to the old prototype clients
(BPAQ/GIS), we have to add JSON, rendering, correct?
As for GIS - have you given any further thoughts on how to handle it
for the GHO? I guess the Swedish/OECD solution may be overkill?
I am still waiting for a decision from Atlassian (they sort of came
back to me a week ago...), and as soon as we get Atlassian Studio, we
should try to consolidate the code there.
Best,
Knut
--
Cheers,
Knut Staring