incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Paul Davis" <paul.joseph.da...@gmail.com>
Subject Re: DAG on couch
Date Fri, 12 Dec 2008 16:58:34 GMT
Did you see the wiki page on storing hierarchical data?

http://wiki.apache.org/couchdb/How_to_store_hierarchical_data

Alot of the ideas on that page require each node to have it's entire
path to the root node which helps in all cases except when you want to
alter a node's path.

Views will never be able to make external calls durring processing.
That'd violate the whole 'insert fancy word' that says map output for
a doc can only depend on the combination of the document + function
source.

On Fri, Dec 12, 2008 at 11:16 AM, Ayende Rahien <ayende@ayende.com> wrote:
> You need a more complex view processor, I think. One that can make
> additional requests during view processing.
>
> On Fri, Dec 12, 2008 at 11:12 AM, Adam Wolff <awolff@gmail.com> wrote:
>
>> Hi everyone,
>> I'm just getting started with CouchDB and I'm really excited about it. I'm
>> working with a directed acyclic graph, and I'd like to be able to
>> efficiently
>> retrieve graph chunks with a single query.
>>
>> So conceptually, my data looks like this
>>
>> doc1: {content: ..., ancestor: null}
>> doc2: {content: ..., ancestor: "doc1"}
>> doc3: {content: ..., ancestor: "doc2"}
>> doc4: {content: ..., ancestor: "doc3"}
>> doc5: {content: ..., ancestor: "doc3"}
>>
>>
>> Long runs of singly linked ancestors will be common; branches (e.g. after
>> doc3
>> in the example above) will be less common but not rare.
>>
>> I get how you can use couch to retrieve a primary key and foreign key at
>> the
>> same time (through view collation) but I'm wondering if there's a clever
>> way
>> to
>> do this recursively.
>>
>> If I could pull depths of 5 nodes or so at a time, that'd be fine. I can
>> always
>> go back for more if I need it. I can imagine a few hacky solutions, like
>> retaining a list of ancestors, but that it makes it pretty much impossible
>> to
>> reorder the graph.
>>
>> Is there a clever solution I'm missing?
>>
>> Thanks in advance,
>> A
>>
>

Mime
View raw message