CODEWITHSUNDEEP
couchdb
Zach Smith
Zach Smith

Reputation: 8961

Is a multi-way join at all possible using CouchDB views?

JOINING data in a 'sql-esque' way is possible within MapReduce as mentioned in many places - such as these Standford course notes, and this MSc thesis, and a plagiarized copy of that thesis published by http://ijniet.org :p.

I have 3 Entities:

  1. A
  2. B
  3. C

A and B can be joined on column ab, and BC can be joined on column bc. I think that I can achieve this via cascading MapReduce tasks. But that I CAN'T do that in CouchDB in any way, whether that is with Map functions, reduce functions, list functions (that consume a MapReduce view), or a combination of these...

In CouchDB you could output documents of different entities with the same join key. But as far as I can tell there is no possible map function that allows for grouping multiple entities by either the same key, or a key range?

This is specifically for JOINS where Foreign key columns are different for each join.

Upvotes: 1

Views: 41

Answers (1)

Alexis Côté
Alexis Côté

Reputation: 3690

You can't do cascade joins in CouchDB.

The only joins that you can do are one-to-many joins.

It consists of a view that emits the linked documents as a value:

function(doc){
    emit([doc._id,0]);
    emit([doc.ab,1],{_id:doc.ab});
}

Then you query with ?include_docs=true

You would get something like this:

{rows:[
    {key:["my_main_doc_id",0],doc:{my main document...}},
    {key:["my_main_doc_id",1],doc:{child doc...}}
]}

See this documentation

Upvotes: 1

Related Questions