【Hive WARN】Column order_discount had to be cast to a less precise type in Hive

妖狐艹你老母 2022-05-22 07:25 110阅读 0赞

最近无聊，随表拷了一个业务数据库中的订单明细表，自己查看了一下导表的日志，发现一直在报一个异常，而且还不少。

具体日志如下：

18/06/19 13:41:56 INFO manager.SqlManager: Executing SQL statement: SELECT t.* FROM `order_info` AS t LIMIT 1
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column logistics_fee had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column order_discount_amount had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column order_amount had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column payment_date had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column send_date had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column recipient_date had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column refund_time had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column creation_date had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column last_update_date had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column actual_amount had to be cast to a less precise type in Hive
    18/06/19 13:41:56 WARN hive.TableDefWriter: Column audit_time had to be cast to a less precise type in Hive
    18/06/19 13:41:56 INFO hive.HiveImport: Loading uploaded data into Hive

虽然，不影响hive表中的数据内容，但是有很多这种类似的报错，感到很不安。

找了很多现在的网上资源，发现大多数都会报这个错误，但是都给略过去了，都没有在意。

但是在我帅气的英文水平下，发现这是个**类型转换，**没错就是个类型转换

我们都知道，Hive是基于JAVA实现的数据建表，数据类型都是JAVA中的数据类型，但是在mysql中，有个数据库类型是小数，而且能指定小数点后几位，前几位的decimal的类型，主要就因为这个类型无法和JAVA同步，在导表的过程中，默认就直接将数据类型转换成了double，但是数据长度是不用的，导致了数据精度不准确。虽然不影响主要计算，但是如果高精度的信息，可能会影像计算结果。

请大家悉知，以上为工作生活中的一点点小收获，分享给大家。