从DynamoDB在Hive创建外部表报错‘Unsupported Hive type: int’求助
Hive创建DynamoDB外部表报错:Unsupported Hive type: int
我尝试通过以下SQL在Hive中基于DynamoDB表创建外部表StoreData:
create external table StoreData(rowid INT, order_priority STRING, discount FLOAT, unit_price FLOAT, shipping_cost FLOAT, customer_id INT, customer_name STRING, ship_mode STRING, customer_segment STRING, product_category STRING, product_subcategory STRING, product_container STRING, product_name STRING, product_base_margin FLOAT, country STRING, region STRING, state_or_province STRING, city STRING, postal_code INT, order_date STRING, ship_date STRING, profit DOUBLE, quantity_ordered_new INT, sales FLOAT, order_id INT) stored by 'org.apache.hadoop.hive.dynamodb.DynamoDBStorageHandler' TBLPROPERTIES ("dynamodb.table.name" = "bigdataassignment1", "dynamodb.column.mapping" = "rowid:`Row ID`,order_priority:`Order Priority`, discount:Discount,unit_price:`Unit Price`,shipping_cost:`Shipping Cost`, customer_id:`Customer ID`,customer_name:`Customer Name`,ship_mode:`Ship Mode`, customer_segment:`Customer Segment`,product_category:`Product Category`, product_subcategory:`Product Sub-Category`,product_container:`Product Container`, product_name:`Product Name`,product_base_margin:`Product Base Margin`, country:Country,region:Region,state_or_province:`State or Province`,city:City, postal_code:`Postal Code`,order_date:`Order Date`,ship_date:`Ship Date`, profit:Profit,quantity_ordered_new:`Quantity ordered new`,sales:Sales,order_id:`Order ID");
该表结构在Redshift中可正常使用,但在Hive执行时触发错误:Unsupported Hive type: int。我确认INT/INTEGER是Hive默认支持类型,且未找到相关错误案例,寻求解决办法。
解决方法
1. 替换INT类型为BIGINT
问题根源在于org.apache.hadoop.hive.dynamodb.DynamoDBStorageHandler对Hive 32位整数类型INT存在兼容性问题——DynamoDB的数值类型(N)默认适配更大范围的整数,部分handler版本未兼容Hive的INT。将所有INT字段替换为BIGINT即可解决:
修改后的核心SQL片段:
create external table StoreData( rowid BIGINT, order_priority STRING, discount FLOAT, unit_price FLOAT, shipping_cost FLOAT, customer_id BIGINT, -- 其余字段按原有逻辑保留,仅将所有INT替换为BIGINT postal_code BIGINT, quantity_ordered_new BIGINT, order_id BIGINT ) -- 存储handler及表属性部分保持不变
2. 升级DynamoDB存储Handler版本
若替换类型后仍报错,需检查hive-dynamodb-handler的版本是否与当前Hive版本兼容。旧版本的handler可能存在类型映射缺陷,建议升级到与Hive版本匹配的最新稳定版。
3. 验证DynamoDB字段类型
确认DynamoDB表中对应Hive整数字段的属性类型为N(数值类型)。如果DynamoDB中是用字符串存储的数字,需将Hive表对应字段改为STRING,后续可通过CAST函数转换为数值类型。
内容的提问来源于stack exchange,提问作者Andriy Sultanov
相关产品推荐
相关产品推荐

