1.3.0 - aggregate functions do not work on embedded documents collection

[dm-ordb]

Versions:
-1.3.0 release
-OrientDB 1.4.0-SNAPSHOT 20130227-0214

Hi, could you try the following sequence:

create class collection

insert into collection set data = [{"size": 0}, {"size": 0}, {"size": 30}]

select
data.size,
data.size(),
min(data.size),
max(data.size),
sum(data.size),
avg(data.size)
from collection

returned:
{
"result": [{
"@type": "d", "@Rid": "#-2:0", "@Version": 0,
"data": [0, 30],
"data2": 3,
"sum": 30,
"avg": 15
}
]
}

expected:
{
"result": [{
"@type": "d", "@Rid": "#-2:0", "@Version": 0,
"data": [0, 0, 30],
"data2": 3,
"min": 0,
"max": 30
"sum": 30,
"avg": 10
}
]
}

It seems the aggregation is made on the set of values (not the list) .

[lvca]

I don't know if it's right the current behavior or not, but try this:

select
list(data.size),
data.size(),
min(data.size),
max(data.size),
sum(data.size),
avg(data.size)

from collection

[dm-ordb]

Luca,

running the following query (added list function as you suggested) does not change anything:

select
list(data.size),
data.size(),
min( list(data.size)),
max( list(data.size)),
sum( list(data.size)),
avg(list(data.size))
from collection

The result is the same and it is actually inconsistent.

On the input we have collection of documents:
[{"size": 0}, {"size": 0}, {"size": 30}]
So the query's task is to aggregate the following collection of numbers [0, 0, 30]

And the result says:
"data": [0, 30], -> 1 element form the collection missed
"data2": 3, -> size of the collection =3 (OK)
"sum": 30, -> OK
"avg": 15 -> incorrect; should be =10 (30/3)
min, max are not calculated at all -> incorrect

It is a bug from my point of view.

Regards
Dan.

[lvca]

So the data is not present in database?

[dm-ordb]

Luca,

I am sorry for inconvenience. I try to explain it more clearly.

If you prepare data:
1.
create class collection
insert into collection set data = [{"size": 0}, {"size": 0}, {"size": 30}, {"size": 50}, {"size": 50}]

and run this query:
2.
select from collection

you get the the correct content in DB:

{"result":[{"@type":"d","@Rid":"#11:0","@Version":0,"@Class":"collection",
"data":[{"size":0},{"size":0},{"size":30},{"size":50},{"size":50}]
}]}

1. But if you try to calculate aggregates on "size" attribute from that collection [[{"size":0},{"size":0},{"size":30},{"size":50},{"size":50}]
   you get incorrect result.

select
data.size as collection_content,
data.size() as collection_size,
min(data.size) as collection_min,
max(data.size) as collection_max,
sum(data.size) as collection_sum,
avg(data.size) as collection_avg
from collection

{"result":[{"@type":"d","@Rid":"#-2:0","@Version":0,
"collection_content":[0,50,30]-> incorrect: should be [0,0, 30, 50, 50] as in the source data
"collection_size":5,-> OK
"collection_sum":80, -> incorrect: should be =130
"collection_avg":26} -> OK
]}

-if you try with [0, 0, 30, 50]
{"result":[{"@type":"d","@Rid":"#-2:0","@Version":0,
"collection_content":[0,50,30], -> incorrect: should be [0, 0, 30, 50] as in the source data
"collection_size":4,-> OK:
"collection_sum":80,-> OK:
"collection_avg":26}-> incorrect; should be =20 (80/4)
]}

-if you try with [0,0,30]
{"result":[{"@type":"d","@Rid":"#-2:0","@Version":0,
"collection_content":[0,30], -> incorrect: should be [0,0, 30] as in the source data
"collection_size":3, -> OK:
"collection_sum":30, -> OK
"collection_avg":15 -> incorrect; should be =10 (30/3)
}
]
}

You can note the following:

• min max aggregates have not been calculated.
• avg, sum result depends on input and sometimes is calculated incorrectly

Regards
Dan

[luigidellaquila]

this is fixed in develop branch.
I just added a test case.

[red-0ne]

Is this merged to 2.2 branch ?