问题
I am trying to migrate data from postgresql server to elasticsearch. The postgres data is in JSONB format. When I am starting the river, I am getting the below error.
[INFO ][logstash.agent ] Successfully started Logstash API endpoint {:port=>9600} [2019-01-07T14:22:34,625][INFO ][logstash.inputs.jdbc ] (0.128981s) SELECT to_json(details) from inventory.retailer_products1 limit 1 [2019-01-07T14:22:35,099][WARN ][logstash.inputs.jdbc ] Exception when executing JDBC query {:exception=>#<Sequel::DatabaseError: Java::OrgLogstash::MissingConverterException: Missing Converter handling for full class name=org.postgresql.util.PGobject, simple name=PGobject>} [2019-01-07T14:22:36,568][INFO ][logstash.pipeline ] Pipeline has terminated {:pipeline_id=>"main", :thread=>"#<Thread:0x6067806f run>"}
I think the logstash is not able to identify the JSON data type.
Below is my logstash conf file
input {
jdbc {
jdbc_connection_string => "jdbc:postgresql://localhost:5432/mydb"
jdbc_user => "postgres"
jdbc_password => "password"
jdbc_validate_connection => true
jdbc_driver_library => "/home/dell5/Downloads/postgresql-9.4.1208.jar"
jdbc_driver_class => "org.postgresql.Driver"
statement => "SELECT to_json(details) from inventory.retailer_products1 limit 1"
}
}
filter{
json{
source => "to_json"
}
}
output {
elasticsearch {
index => "products-retailer"
document_type => "mapping-retailer"
hosts => "localhost"
}
stdout{}
}
The mapping I have defined for this is as below
{
"products-retailer": {
"mappings": {
"mapping-retailer": {
"dynamic": "false",
"properties": {
"category": {
"type": "keyword"
},
"id": {
"type": "keyword"
},
"products": {
"type": "nested",
"properties": {
"barcode": {
"type": "text"
},
"batchno": {
"type": "text"
},
"desc": {
"type": "text"
},
"expirydate": {
"type": "date",
"format": "YYYY-MM-DD"
},
"imageurl": {
"type": "text"
},
"manufaturedate": {
"type": "date",
"format": "YYYY-MM-DD"
},
"mrp": {
"type": "text"
},
"name": {
"type": "text",
"fields": {
"ngrams": {
"type": "text",
"analyzer": "autocomplete"
}
}
},
"openingstock": {
"type": "text"
},
"price": {
"type": "text"
},
"purchaseprice": {
"type": "text"
},
"sku": {
"type": "text"
},
"unit": {
"type": "text"
}
}
},
"retailerid": {
"type": "keyword"
},
"subcategory": {
"type": "keyword"
}
}
}
}
}
}
The sample data in postgres column is below. It has nested json that I have defined in the mapping of elasticsearch.
{
"id": "",
"Category": "Bread and Biscuits",
"products": {
"MRP": "45",
"SKU": "BREAD-1",
"Desc": "Brown Bread",
"Name": "Brown Bread",
"Unit": "Packets",
"Brand": "Britannia",
"Price": "40",
"BarCode": "1234567890",
"BatchNo": "456789",
"ImageUrl": "buscuits.jpeg",
"ExpiryDate": "2019-06-01",
"OpeningStock": "56789",
"PurchasePrice": "30",
"ManufactureDate": "2018-11-01"
},
"RetailerId": "1",
"SubCategory": "Bread"
}
Please suggest what am I missing here and if this is the right way to do it.
I am using Elasticsearch 6.5.1. PostgreSQL 9.5.
回答1:
PGObject does not have capability to convert the json came from to_json
method. use internal casting for converting jsonobject to text like this.
SELECT to_json(details)::text from inventory.retailer_products1 limit 1.
Now you can parse the json string in logstash.
回答2:
I have been hitting the same error today and it appears that logstash is unable to convert Postgres json or jsonb data fields (PgObject).
I cast my object fields into TEXT datatype and it stopped yelling at me and started ingesting data. As to whether it is indexing correctly remains to be seen.
I will update my response when i have a better idea if this was the correct approach.
来源:https://stackoverflow.com/questions/54085353/not-able-to-insert-json-from-postgresql-to-elasticsearch-getting-error-excep