ElasticSearch aggregate information grouped by month - elasticsearch

Question

I am working on an ElasticSearch query that should give me back all documents where a date field is now-1 year ago and then group them all by month (giving me total count for each month), but am failing on writing this query.
This is what I have:
{
"query": {
"bool": {
"must": [
{
"terms": {
"account_id": [
1
]
}
}
]
}
},
"aggs": {
"growth": {
"date_range": {
"field": "member_since",
"format": "YYYY-MM-DD",
"ranges": [
{
"to": "now-1Y/M"
},
{
"from": "now-1Y/M"
}
]
}
}
},
"size": 100
}
I am running the query like so:
POST https://my-es-cluster-url.com, but I keep getting this error:
{
"error": {
"root_cause": [
{
"type": "parse_exception",
"reason": "unit [Y] not supported for date math [-1Y/M]"
}
],
"type": "search_phase_execution_exception",
"reason": "all shards failed",
"phase": "query_fetch",
"grouped": true,
"failed_shards": [
{
"shard": 0,
"index": "index0",
"node": "MkypXlGdQamAplca1JIgZQ",
"reason": {
"type": "parse_exception",
"reason": "unit [Y] not supported for date math [-1Y/M]"
}
}
]
},
"status": 400
}

Answer 1

I reproduced your problem with a simpler query on my data.
"query": {
"range": {
"recent_buy_transaction": {
"from": "now-1Y"
}
}
}
I get the same error.
ElasticSearchParseException[unit [Y] not supported for date math [-1Y]]
However, using small y fixes the problem. Hence, you should try using:
now-1y/M
From the documentation
The supported time units are: y (year), M (month), w (week), d (day), h (hour), m (minute), and s (second).