Re: [RavenDB] MapReduce over Another MapReduce

[Fitzchak Yitzchaki]

This is not supported.

See this answer on StackOverflow for more details.

What is that you're trying to achieve?

You can save those aggregations as documents.

On Sun, Aug 12, 2012 at 2:56 PM, SwatB <bhardwa...@hotmail.com> wrote:

Hello,

I plan to use ravendb for time-series data.

Is it possible to chain output of one map-reduce to another map-reduce.

Say for example - I create a map-reduce to create daily aggregation.

Then I use another map-reduce on the previous results to get fortnightly aggregation.

In both of the cases, I will like to have incremental map-reduce. 

If there is an example of this kind of scenario please do point to me.

regards

[Oren Eini (Ayende Rahien)]

No, create two indexes for this

[SwatB]

Hello,

Thanks for your reply

I read your answer, there is something like out in MondoDB, is that what you mean by saving  aggregations as documents. (a new collection)

What if there is another change in the base data, is there a mechanism by which I can do upsert , I mean do the incremental thing

[Fitzchak Yitzchaki]

I assumed that the data won't be changed. If it may be changed, you'll need to go with another solution.

[Oren Eini (Ayende Rahien)]

Or he can use the indexed properties option

[SwatB]

Hello Oren,

What do you mean I can use " indexed properties option "

regards

[Oren Eini (Ayende Rahien)]

http://issues.hibernatingrhinos.com/issue/RavenDB-245