fix(bigquery): update the API

#### bigquery:v2
The following keys were changed:
- schemas.QueryRequest.properties.requestId.description

discovery/bigquery-v2.json

@@ -1519,7 +1519,7 @@
       }
     }
   },
-  "revision": "20200523",
+  "revision": "20200603",
   "rootUrl": "https://bigquery.googleapis.com/",
   "schemas": {
     "AggregateClassificationMetrics": {
@@ -4358,7 +4358,7 @@
           "type": "array"
         },
         "requestId": {
-          "description": "A unique user provided identifier to ensure idempotent behavior for queries. Note that this is different from the job_id. It has the following properties: 1. It is case-sensitive, limited to up to 36 ASCII characters. A UUID is recommended. 2. Read only queries can ignore this token since they are nullipotent by definition. 3. When a duplicate mutating query request is detected (i.e. having the same request_id as an earlier query), it returns: a. the results of the mutation if it completes successfully within the timeout. b. the running operation if it is still in progress at the end of the timeout. 4. Its lifetime is limited to 15 minutes. In other words, if two requests are sent with the same request_id, but more than 15 minutes apart, idempotency is not guaranteed.",
+          "description": "A unique user provided identifier to ensure idempotent behavior for queries. Note that this is different from the job_id. It has the following properties: 1. It is case-sensitive, limited to up to 36 ASCII characters. A UUID is recommended. 2. Read only queries can ignore this token since they are nullipotent by definition. 3. For the purposes of idempotency ensured by the request_id, a request is considered duplicate of another only if they have the same request_id and are actually duplicates. When determining whether a request is a duplicate of the previous request, all parameters in the request that may affect the behavior are considered. For example, query, connection_properties, query_parameters, use_legacy_sql are parameters that affect the result and are considered when determining whether a request is a duplicate, but properties like timeout_ms don't affect the result and are thus not considered. Dry run query requests are never considered duplicate of another request. 4. When a duplicate mutating query request is detected, it returns: a. the results of the mutation if it completes successfully within the timeout. b. the running operation if it is still in progress at the end of the timeout. 5. Its lifetime is limited to 15 minutes. In other words, if two requests are sent with the same request_id, but more than 15 minutes apart, idempotency is not guaranteed.",
           "type": "string"
         },
         "timeoutMs": {

src/apis/bigquery/v2.ts

@@ -2069,7 +2069,7 @@ export namespace bigquery_v2 {
      */
     queryParameters?: Schema$QueryParameter[];
     /**
-     * A unique user provided identifier to ensure idempotent behavior for queries. Note that this is different from the job_id. It has the following properties: 1. It is case-sensitive, limited to up to 36 ASCII characters. A UUID is recommended. 2. Read only queries can ignore this token since they are nullipotent by definition. 3. When a duplicate mutating query request is detected (i.e. having the same request_id as an earlier query), it returns: a. the results of the mutation if it completes successfully within the timeout. b. the running operation if it is still in progress at the end of the timeout. 4. Its lifetime is limited to 15 minutes. In other words, if two requests are sent with the same request_id, but more than 15 minutes apart, idempotency is not guaranteed.
+     * A unique user provided identifier to ensure idempotent behavior for queries. Note that this is different from the job_id. It has the following properties: 1. It is case-sensitive, limited to up to 36 ASCII characters. A UUID is recommended. 2. Read only queries can ignore this token since they are nullipotent by definition. 3. For the purposes of idempotency ensured by the request_id, a request is considered duplicate of another only if they have the same request_id and are actually duplicates. When determining whether a request is a duplicate of the previous request, all parameters in the request that may affect the behavior are considered. For example, query, connection_properties, query_parameters, use_legacy_sql are parameters that affect the result and are considered when determining whether a request is a duplicate, but properties like timeout_ms don't affect the result and are thus not considered. Dry run query requests are never considered duplicate of another request. 4. When a duplicate mutating query request is detected, it returns: a. the results of the mutation if it completes successfully within the timeout. b. the running operation if it is still in progress at the end of the timeout. 5. Its lifetime is limited to 15 minutes. In other words, if two requests are sent with the same request_id, but more than 15 minutes apart, idempotency is not guaranteed.
      */
     requestId?: string | null;
     /**