CouchDB - aggregate multiple values in the reduce step - mapreduce

Question

I am trying to aggregate multiple values emitted from my map in a custom reduce function( rereduce enabled) but I get null values in the result except for few datapoints where I see 0s., Any help is much appreciated. Thanks
My data emitted is in this format:
Key:"dateX" Values:"{'a':435, 'b':5645}"
Key:"dateX" Values:"{'a':8451, 'b':9245}"
Key:"dateX" Values:"{'a':352, 'b':450}"
Key:"dateY" Values:"{'a':5675, 'b':1245}"
Key:"dateY" Values:"{'a':4455, 'b':620}"
I want to aggregate/sum the values of both a & b for dateX and dateY, my map-reduce is:
"map": "function(doc){emit(doc.logDate, {'a': doc.a, 'b': doc.b} );}",
"reduce": "function(key, values, rereduce) {
var total = {tA:0, tB:0};
if(rereduce){
for(i=0; i<values.length; i++)
{
total.tA += values[i].a;
total.tB += values[i].b;
}
return total;
}
total.tA = sum(values.a);
total.tB = sum(values.b);
return total; }"
------------------------
Results:
dateX {tA: 0, tB: 0}
dateY {tA: null, tB: null}
dateZ {tA: null, tB: null}
.
.
.

Answer 1

First of all function sum expects an array as an argument instead of values.a which is undefined, so you can try
total.tA = sum(values.map(function (value) {return value.a}));
total.tB = sum(values.map(function (value) {return value.b}));
Or something like
values.reduce(function (t, value) {
t.tA += value.a;
t.tB += value.b;
return t;
}, total);
Also, when rereduce flag is enabled means that the reduce function works with already reduced values, so
for (i=0; i < values.length; i++) {
total.tA += values[i].tA; //not a
total.tB += values[i].tB; //not b
}
Also, you should add else to prevent double execution in rereduce case
if (rereduce) {
//
} else {
//
}
http://guide.couchdb.org/editions/1/en/views.html#reduce