# Inconsistent Aggs When Using Nested Docs
**URL:** https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519
**Category:** Elasticsearch
**Created:** [February 26, 2017, 4:09am UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519 "2017-02-26T04:09:34Z")
**Posts on this page:** 8
**Page:** 1
### Author:  [@neoform](https://discuss.elastic.co/u/neoform)
#### Post date: [February 26, 2017, 4:09am UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/1 "2017-02-26T04:09:34Z")
**Elasticsearch version** : 5.2.1
**JVM version** : 1.8.0\_121
**OS version** : Debian 8
I'm using a single shard index (no replicas).
For some reason my query's aggs are giving me numbers that are either exactly correct, or greater than the query's doc count.
Since I'm using an index that uses multiple levels of nested docs, I have to use an "reverse\_nested" agg to get the doc count when filtering by a nested doc value.
Any ideas why this would happen? (I can provide a copy of the full db if needed, this isn't prod/sensitive info).
#Query
```javascript
{
"size": 0,
"_source": false,
"query": {
"bool": {
"filter": [
{
"nested": {
"path": "product",
"query": {
"bool": {
"filter": [
{
"term": {
"product.status_id": 1
}
},
{
"nested": {
"path": "product.brand",
"query": {
"bool": {
"filter": [
{
"term": {
"product.brand.id": 23107 //<-- additional product.brand.id filter
}
}
]
}
}
}
}
]
}
}
}
}
]
}
},
"aggs": {
"brands": {
"global": {}, //<---- global scope
"aggs": {
"brands": {
"filter": {
"bool": {
"filter": [
{
"nested": {
"path": "product",
"query": {
"bool": {
"filter": [
{
"term": {
"product.status_id": 1 //<-- same filter as above minus product.brand.id
}
}
]
}
}
}
}
]
}
},
"aggs": {
"brands": {
"nested": {
"path": "product.brand"
},
"aggs": {
"brands": {
"terms": {
"field": "product.brand.id",
"size": 5
},
"aggs": {
"brands": {
"reverse_nested": {} //<-- get root doc count
}
}
}
}
}
}
}
}
}
}
}
```
#Results:
```auto
{
"took": 226,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"failed": 0
},
"hits": {
"total": 2890, // <-- doc count
"max_score": 0,
"hits": []
},
"aggregations": {
"brands": {
"doc_count": 1354323,
"brands": {
"doc_count": 442506,
"brands": {
"doc_count": 500478,
"brands": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 465820,
"buckets": [
{
"key": 11317,
"doc_count": 9877,
"brands": {
"doc_count": 9877
}
},
{
"key": 28221,
"doc_count": 6805,
"brands": {
"doc_count": 6805
}
},
{
"key": 40330,
"doc_count": 6243,
"brands": {
"doc_count": 6243
}
},
{
"key": 23107, //<--- same product.brand.id as used in main query
"doc_count": 6132,
"brands": {
"doc_count": 2948 //<--- main query doc count is 2890
}
},
{
"key": 23083,
"doc_count": 5601,
"brands": {
"doc_count": 5599
}
}
]
}
}
}
}
}
}
```
---
### Author:  [@jkuang](https://discuss.elastic.co/u/jkuang)
#### Post date: [February 27, 2017, 7:10pm UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/2 "2017-02-27T19:10:59Z")
The difference is due to your query. First your query filter by product.status\_id = 1 and then [product.brand.id](http://product.brand.id) = 23107 . This result in 2890 docs returned. In your aggs you only specified filter product.status\_id = 1.
---
### Author:  [@neoform](https://discuss.elastic.co/u/neoform)
#### Post date: [February 27, 2017, 9:48pm UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/3 "2017-02-27T21:48:40Z")
Right, but I'm doing an agg on the [product.brand.id](http://product.brand.id) which creates buckets with counts, then the corresponding bucket for [product.brand.id](http://product.brand.id) = 23107 shows a count of 2948 even though I rejoin the agg to the doc root with 'reverse\_nested'.
---
### Author:  [@jkuang](https://discuss.elastic.co/u/jkuang)
#### Post date: [February 28, 2017, 12:54am UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/4 "2017-02-28T00:54:24Z")
Can you try removing the product.brand agg?
From:
```auto
"aggs": {
"brands": {
"nested": {
"path": "product.brand"
},
"aggs": {
"brands": {
"terms": {
"field": "product.brand.id",
"size": 5
},
```
To
```
"aggs": {
"brands": {
"terms": {
"field": "product.brand.id",
"size": 5
},
```
---
### Author:  [@neoform](https://discuss.elastic.co/u/neoform)
#### Post date: [February 28, 2017, 1:13am UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/5 "2017-02-28T01:13:54Z")
That query wouldn't work since we're dealing with nested documents. The agg would return nothing.
---
### Author:  [@jkuang](https://discuss.elastic.co/u/jkuang)
#### Post date: [February 28, 2017, 4:08pm UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/6 "2017-02-28T16:08:21Z")
I'm able to get my sample to work and it's nested doc. It would be easier if you can provide some sample data and your index mapping.
```auto
My example:
......
......
"aggs": {
"all": {
"global": {},
"aggs": {
"price": {
"filter": {
"bool": {
"filter": [
{
"nested": {
"path": "user",
"query": {
"bool": {
"filter": [
{
"term": {
"user.price": 1 //<-- same filter as above minus cost.id
}
}
]
}
}
}
}
]
}
},
"aggs": {
"brands": {
"nested": {
"path": "cost"
},
"aggs": {
"brands": {
"reverse_nested": {} //<-- get root doc count
}
}
}
}
}}}}
```
```auto
Output:
{
"took": 3,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 3,
"max_score": 0,
"hits": [
.....
....
"aggregations": {
"all": {
"doc_count": 6,
"price": {
"doc_count": 4,
"brands": {
"doc_count": 7,
"brands": {
"doc_count": 3
}
}
}
}
}
```
---
### Author:  [@neoform](https://discuss.elastic.co/u/neoform)
#### Post date: [March 1, 2017, 4:55am UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/7 "2017-03-01T04:55:38Z")
An ElasticSearch contributor solved my issue here: [https://github.com/elastic/elasticsearch/issues/23319](https://github.com/elastic/elasticsearch/issues/23319)
---
### Author:  [@system](https://discuss.elastic.co/u/system)
#### Post date: [March 29, 2017, 4:56am UTC](https://discuss.elastic.co/t/inconsistent-aggs-when-using-nested-docs/76519/8 "2017-03-29T04:56:23Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.