Decay function across multiple indices where decay field name might not exist - elasticsearch

Question

I'm trying to use a decay function across multiple indices. The issue I'm having is that some of those indices do not have the field that the decay function uses so I'm getting the following error.
{"type":"parsing_exception","reason":"unknown field [start_date]","line":1,"col":0},{"type":"parsing_exception","reason":"unknown field [match_date]","line":1,"col":0}
I can provide any additional information that would be helpful.
The documentation shows that there is a filter option that I attempted to use however it is still tripping up when not finding the field.
I'm using the php flavor of elasticsearch.
'query' => [
"function_score" => [
'query' => [
"bool" => [
"must" => [
"multi_match" => [
"query" => $search,
"fields" => [
"match_web_name",
"shooter_result_name",
"name",
"post_match_name",
"club_name"
]
]
],
"should" => [
[
"term" => [
"private" => 0
],
]
]
]
],
"functions" => [
[
"filter" => [ "exists" => ["field" => "start_date"]],
"linear" => [
"start_date" => [
"origin" => 'now',
"scale" => $scale
]
]
],
[
"filter" => [ "exists" => ["field" => "match_date"]],
"linear" => [
"match_date" => [
"origin" => 'now',
"scale" => $scale
]
]
]
]
]
]
Here is the mapping for an index that does not have the start_date
"club": {
"properties": {
"zip": {
"type": "text",
"fields": {
"raw": {
"type": "keyword"
}
}
},
"website": {
"type": "text",
"fields": {
"raw": {
"type": "keyword"
}
}
},
"city": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
},
"raw": {
"type": "keyword"
}
}
},
"about": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
}
}
},
"tagline": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
}
}
},
"location": {
"ignore_malformed": true,
"type": "geo_point"
},
"id": {
"type": "integer"
},
"state": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
},
"raw": {
"type": "keyword"
}
}
},
"club_name": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
},
"raw": {
"type": "keyword"
}
}
},
"deleted_at": {
"format": "yyyy-MM-dd HH:mm:ss",
"type": "date"
},
"slug": {
"type": "keyword"
}
}
Here is a mapping where an index does contain the start_date
"match": {
"properties": {
"end_date": {
"null_value": "1990-01-01 12:12:12",
"format": "yyyy-MM-dd HH:mm:ss",
"type": "date"
},
"zip": {
"type": "text",
"fields": {
"raw": {
"type": "keyword"
}
}
},
"private": {
"type": "boolean"
},
"reg_close_date": {
"null_value": "1990-01-01 12:12:12",
"format": "yyyy-MM-dd HH:mm:ss",
"type": "date"
},
"city": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
},
"raw": {
"type": "keyword"
}
}
},
"level": {
"type": "keyword"
},
"description": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
}
}
},
"max_shooters": {
"type": "integer"
},
"type": {
"type": "keyword"
},
"deleted_at": {
"format": "yyyy-MM-dd HH:mm:ss",
"ignore_malformed": true,
"type": "date"
},
"reg_open_date": {
"null_value": "1990-01-01 12:12:12",
"format": "yyyy-MM-dd HH:mm:ss",
"type": "date"
},
"score_type": {
"type": "keyword"
},
"club_id": {
"type": "integer"
},
"match_web_name": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
},
"raw": {
"type": "keyword"
}
}
},
"location": {
"ignore_malformed": true,
"type": "geo_point"
},
"id": {
"type": "integer"
},
"state": {
"type": "text",
"fields": {
"en": {
"analyzer": "english",
"type": "text"
},
"raw": {
"type": "keyword"
}
}
},
"prices": {
"type": "text",
"fields": {
"raw": {
"type": "keyword"
}
}
},
"slug": {
"type": "keyword"
},
"start_date": {
"null_value": "1990-01-01 12:12:12",
"format": "yyyy-MM-dd HH:mm:ss",
"type": "date"
},
"status": {
"type": "keyword"
}
}