CDA - DataAccess - Settings-Block + Output-Lines

[neckit]

I just study the possibilities of CDA and I'm thrilled

I've read the CDA-Documentation and all is clear - except two things, i found in the samples:

1. Settings-Block under DataAccess
OK - i can override the defaults for the output-type to e.g. CSV.
But what are with the other properties (includeTotals; aggregator; and so on)
I can't find any documentation. What options do I have in the Settings-block?

2. In the Output-Area of the jndi-sample-file i found following lines:

<Output indexes="1,0,2,3"/>
<Output id="2" indexes="0,1,3"/>
<Output id="3" indexes="1,0,2"/>

The first line is clear. But what is the meaning of the other lines?

greetings
Ralf

[pmalves]

We need to document that. I actually dont know the answer, xpandit guys commited that feature.

Basic idea is that you can define multiple outputs, and at runtime you can choose the output format you want

[flamierd]

Actually I was looking for just this feature. At this point I was considering creating multiple Compound "union" blocks with some dummy data in order to generate the necessary variable outputs I want for my charts. In my case it's that I have hourly revenue, transaction counts, response times, etc I want to display in different charts on the same dashboard. Hitting the database 4 different times to do pulls has massive overhead.


Of course there is no support for &outputIndexId=2 parameter in the cda datasource component of cde...<hint>

May end up for now experimenting with the unions.

Unions don't apparently support the <output> tag which is odd since the "join" apparently does.

Update: The "join" almost works as expected. In the CDA previewer it works perfectly, however feeding the new DataAccess to the exact same chart causes some sort of error, perhaps the numeric has been converted to a string in the transform? Investigating.

Update 2: The problem was a bit mysterious but I finally worked it out. When there is no match on a "join" the system seems to be generating an additional row of nulls. This isn't what I expected from an outer join in Oracle, it may be ANSI standard but otherwise I think it's a bug. So in any case I simply had the system return a value and created the duplicate value in my table and had them join on that value. Works good. Suitable work around until OutputIndex is supported by CDE.

[flamierd]

Bravo on the inclusion of the OutputIndexId options in the CDE. Mucho grande easier for me to flow through my data to multiple components or even multiple dashes using caches.