hive 导入导出csv文件

1、导入
将CSV文件导入到hive数据库，具体操作如下，
首先在hive中创建自己的table, 并且设置以逗号分隔。

create table if not exists tmp.hb_label_C(
sample_cat string,
reg_mob_md5 string)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ',' STORED AS TEXTFILE;

然后

desc formatted  tmp.hb_label_C

确定tmp.hb_label_C表的分布式文件路径

最后在linux命令行输入

hdfs dfs -put label_C.csv hdfs://nameservice1/user/hive/warehouse/tmp.db/hb_label_C

查看结果

select * from tmp.hb_label_C limit 10;

2、导出

使用beeline重定向

beeline -e "select * from tmp.install_uninstall_pred_res_online_v1 where tail<=9" > hb_torch_feature.csv

同时使用以下方式导出数据带表头

beeline -e "select * from tmp.install_uninstall_pred_res_online_v1 where tail<=9"  |sed '1s/^/md5_phone\n/' > hb_torch_feature.csv

hive-cli 方式

hive -e "set hive.cli.print.header=true;select * from test.jxb_fpd_agr1_all_0905;" |grep -v "WARN" | tr "\t" "," > /jinxb/result/test.csv