Elastic Search Aggregation buckets, buckets by number of records - elasticsearch

Question

I am new to Elastic Search and I'm trying to create a request without a lot of success. Here is the use case:
Let's imagine I have 4 documents, which have an amount field:
[
{
"id": 541436748332,
"amount": 5,
"date": "2017-01-01"
},
{
"id": 6348643512,
"amount": 2,
"date": "2017-03-13"
},
{
"id": 343687432,
"amount": 2,
"date": "2017-03-14"
},
{
"id": 6457866181,
"amount": 7,
"date": "2017-05-21"
}
]
And here is the kind of result I'd like to get:
{
"aggregations": {
"my_aggregation": {
"buckets": [
{
"doc_count": 2,
"sum": 7
},
{
"doc_count": 2,
"sum": 9
}
]
}
}
}
As you can see, I want some kind of histogram, but instead of putting a date interval, I'd like to set a "document" interval. So here, that would be 2 documents per bucket, and the sum of the field amount of those two documents.
Does someone knows if that is even possible? That would also imply to sort the records by date for example, to get the wanted results
EDIT: Some more explanations on the use case:
The real use case is a line graph I'd like to print. But I want to make the X axis the number of sales, and in the Y the total amount $$$ of those sales. And I don't want to print thousands of dot on my graph, I want fewer dots, that's why I was hoping to deal with the buckets and the sums...
The example of response I gave is just the first step I want to achieve, the second step would be to add each field the one that is behind it:
{
"aggregations": {
"my_aggregation": {
"buckets": [
{
"doc_count": 2,
"sum": 7
},
{
"doc_count": 2,
"sum": 16
}
]
}
}
}
(7 = 5 + 2); (16 = 7 (from last result) + 2 + 7);

Answer 1

You can use histogram and sum aggregations, like this:
{
"size": 0,
"aggs": {
"prices": {
"histogram": {
"field": "id",
"interval": 2,
"offset": 1
},
"aggs": {
"total_amount": {
"sum": {
"field": "amount"
}
}
}
}
}
}
(offset 1 is required if you want the first bucket to start at 1 instead of at 0.) Then you'll get a response like this:
{
"aggregations": {
"prices": {
"buckets": [
{
"key": 1,
"doc_count": 2,
"total_amount": {
"value": 7
}
},
{
"key": 3,
"doc_count": 2,
"total_amount": {
"value": 9
}
}
]
}
}
}
Sorting is not required, because the default order is the order you want. However, there's also an order parameter in case you want a different ordering of the buckets.