Hi. Is it possible to use the interval of a date histogram aggregation in calculations of nested aggs ?
For instance, I have a query something like this:
{
"size": 0,
"query":{"filtered":{"query":{"match_all":{}}}}
,"aggs": {
"x.interval": {
"date_histogram": {
"field": "timestamp",
"interval": "1m",
"pre_zone": "+02:00",
"pre_zone_adjust_large_interval": true,
"min_doc_count": 1,
"extended_bounds": {
"min": 1432215551896,
"max": 1432219151896
}
},
"aggs": {
"y.value": {
"sum": {
"field": "to.bytes"
}
}
}
}
}
}
That gives me a result something like this
"aggregations": {
"x.interval": {
"buckets": [
{
"key_as_string": "2015-05-21T12:58:00.000Z",
"key": 1432213080000,
"doc_count": 15132,
"y.value": {
"value": 236036536
}
},
{
"key_as_string": "2015-05-21T14:34:00.000Z",
"key": 1432218840000,
"doc_count": 17430,
"y.value": {
"value": 5360547741
}
},
{
"key_as_string": "2015-05-21T14:35:00.000Z",
"key": 1432218900000,
"doc_count": 61721,
"y.value": {
"value": 5192133328
}
},
{
"key_as_string": "2015-05-21T14:36:00.000Z",
"key": 1432218960000,
"doc_count": 53301,
"y.value": {
"value": 4304589515
}
}
So, the inner buckets "y.value" is the sum of bytes used for a specific date interval (outer date aggregation "x.interval").
But I want to calculate the y.value as "sum(bytes) / interval" , to give me a bytes/second rate?
Thanx in advance.
Pieter