Hello, it seems that the append processor "stringifies" input arrays instead of processing the elements one by one. Could that be? (Using elasticsearch 8.6.2)
Pipeline:
"description": "testing issues with append",
"processors": [
{
"append": {
"field": "copy_to",
"value": [
"{{{from}}}"
],
"ignore_failure": true
}
}
]
Test document:
[
{
"_index": "o",
"_id": "i",
"_source": {
"from": [
2,
3
]
}
}
]
Simulated output:
{
"docs": [
{
"doc": {
"_index": "o",
"_id": "i",
"_version": "-3",
"_source": {
"copy_to": [
"{0=2, 1=3}"
],
"from": [
2,
3
]
},
"_ingest": {
"timestamp": "2023-05-15T16:11:17.691991822Z"
}
}
}
]
}
I'd expect copy_to
to be [2, 3]
, not {0=2, 1=3}
.