Currently, the GraphQLValidation does not check if the row type of the CREATE TABLE for a mutation matches the GraphQL Input type, resulting in runtime errors.
CREATE TABLE DeploymentMetrics (
`deploymentId` STRING,
`service` STRING, -- the service (e.g. flink, postgres, vertx)
`component` STRING, -- component within a service (i.e. task-/job manager)
`taskName` STRING, -- from labels['task_name'] in Flink, or labels['query'] in Vertx
`name` STRING, -- name of the metric
`metric` DOUBLE, -- the value of simple metrics or max for histograms
`bucketValues` ARRAY<DOUBLE NOT NULL>, -- values for histogram buckets (latencies)
`timestamp` TIMESTAMP_LTZ(3) NOT NULL,
WATERMARK FOR `timestamp` AS `timestamp` - INTERVAL '0.001' SECOND
);
will compile with Input type:
input DeploymentMetricsInput {
deploymentId: String
service: String
instanceId: String
task_name: String
name: String
metric: Float
bucketValues: [Float]
timestamp: DateTime
uid: String
}
Currently, the GraphQLValidation does not check if the row type of the CREATE TABLE for a mutation matches the GraphQL Input type, resulting in runtime errors.
will compile with Input type: