hive 实践

1. 创建hive分区表

create table source(
`date` bigint,
event int,
app string,
domain string,
rtype int,
unique_name string,
user_id string,
host_name string,
cluster_info string,
client_isp string,
isp string,
node string,
protocol string,
count bigint,
transcode_count bigint,
transcode_real_kbps bigint,
real_kbps bigint,
in_bytes bigint,
out_bytes bigint,
ntype string,
status_code int,
log_count bigint,
transcode_log_count bigint
)
PARTITIONED BY ( 
  `day` string, 
  `hour` string, 
  `m5` string)
ROW FORMAT SERDE 
  'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe' 
STORED AS INPUTFORMAT 
  'org.apache.hadoop.mapred.TextInputFormat' 
OUTPUTFORMAT 
  'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
LOCATION
  'hdfs://hdfs-ha/apps/hive/warehouse/eag.db/source'
TBLPROPERTIES (
  'last_modified_by'='hive', 
  'last_modified_time'='1585206348', 
  'transient_lastDdlTime'='1585206348');

2.查看hive表的建表语句

show create table tablename;

3.写hive库，需要hive-site.xml配置文件

cat /usr/hdp/2.6.1.0-129/spark2/conf/hive-site.xml

hive.metastore.uris=thrift://xx1:9083,thrift://xx2:9083
在项目的conf新增hive-site.xml，并在启动脚本中配置加载hive-site.xml

   
    
      hive.metastore.uris
      thrift://xx1:9083,thrift://xx2:9083
    
    

4.程序运行报错：

org.apache.spark.sql.AnalysisException: `eag`.`source` requires that the data to be inserted have the same number of columns as the target table: target table has 28 column(s) but the inserted data has 25 column(s), including 0 partition column(s) having constant value(s).;
原因：写入的数据字段与表的字段个数不对应

val hivedf = hiveCtx.read.schema(OFFLINESOURCESCHEMA).json(df)

重新检查schema对应的字段与hive表字段是否一一对应

5.权限问题：

Permission denied: user=spark, access=WRITE, inode="/apps/hive/warehouse/eag.db/source/.hive-staging_hive_2020-03-26_16-33-03_236_3911206415334421369-1":hdfs:hdfs:drwxr-xr-x

解决办法：在启动脚本开头加上export HADOOP_USER_NAME=hdfs