org.apache.kafka.connect.errors.DataException: Invalid Java object for schema with type FLOAT64: class java.lang.Long for field:

[lucasenabit]

please, I had a problem with this field. In my ES i have a scheme with long field, but i have too others fields with same type, but this problem happen just that field.

Exemple mapping ES:

},
"redirectEnd" : {
"type" : "long"
},
"redirectStart" : {
"type" : "long"
},
"requestStart" : {
"type" : "float"
},
"resourceType" : {
"type" : "keyword"
},
"responseEnd" : {
"type" : "float"
},
"responseStart" : {
"type" : "float"
},
"secureConnectionStart" : {
"type" : "float"
},
"startTime" : {
"type" : "float"
},
"transferSize" : {
"type" : "long"
},
"workerStart" : {
"type" : "long"
}

Exemple values ES:

{
"connectEnd" : 4898.4000000059605,
"connectStart" : 4898.4000000059605,
"encodedBodySize" : 21504,
"decodedBodySize" : 140288,
"transferSize" : 8192,
"domainLookupEnd" : 4898.4000000059605,
"domainLookupStart" : 4898.4000000059605,
"duration" : 3224,
"entryType" : "resource",
"fetchStart" : 4898.4000000059605,
"initiatorType" : "script",
"resourceType" : "js",

tks.

[DarioBalinzo]

Hi, thanks for reporting this.

The connector guess the schema looking at the values in the json, and this problem is caused by a field that contains both number with and without digits.

I will schedule a fix for this, since the connector might use the es schema as source of truth for decide data types, instead of just looking the values of the data.

[lucasenabit]

Hi Dario, tks for answer.

so, at the moment in this case, would not have an alternative to this problem?

[DarioBalinzo]

As workaround, try one of the following:

• try to use the json serializer (not avro)
• index data with a .0 decimal digits, also if they are integers
• blacklist the field in order to avoid the error