Hive-SQL练习
通过hive统计一篇文章中WordCount
1. hive创建内部表wc
create table wc ( lie string) row format delimited fields terminated by '\n';
2. 向wc表中导入hdfs文章的内容
load data inpath '/1.txt' overwrite into table wc;
3. 统计词频
select word,count(1) as word_count from (select explode(split(lie2, " ")) as word //将文章的每一行(表中的每条记录),按照空格切割为每个元素 from (select regexp_replace(lie,'[^a-zA-Z0-9]',' ') as lie2 //过滤掉特殊字符,只保留a-zA-Z0-9,其余全部替换为空格 from wc ) t1 ) t where word REGEXP '[a-zA-Z0-9]' //过滤掉空格 group by word order by word_count desc //按照词频排倒序(desc降序,默认asc升序) limit 10;
4. 知识点
regexp_replace(source,reg规则,desc) //字段source内容满足reg规则就替换为desc指定的内容
比如:regexp_replace(123,'^[0-9]',123) //返回12323