用Hue分析马老师的B站视频弹幕----我发现了一个新词

CREATE DATABASE IF NOT EXISTS test;

use  test;

create external table MBGAnalyse (dm_time STRING,send_date STRING,send_month STRING,send_time STRING,text STRING,user_id STRING) row format delimited fields terminated by ',';

load data inpath '/user/hive/warehouse/ababaaba/biliMBG.csv' into table MBGAnalyse;

SELECT word ,count(1) as count FROM 
(SELECT explode(split(regexp_replace(ma.text,'[^A-Za-z0-9\\u4e00-\\u9fa5]',''),' '))as word  FROM mbganalyse ma) W
GROUP BY word
ORDER BY word desc ;