Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hive >> mail # user >> Correct way of using regexserde


Copy link to this message
-
Re: Correct way of using regexserde
Hello,
try to delete  "output.regex" = "%1$s" and store your data in file then
pricise where the file is located
STORED AS TEXTFILE
LOCATION '/......';
and assure that you add the jar "hive-contrib-0.10.0.jar"  in the session
or you have it in all Hadoop tasktrackers
like this:
add jar path_where_is_the_jar_in_hive_lib\hive-contrib-0.9.0.jar ;

Good luck
2013/7/1 Mohammad Tariq <[EMAIL PROTECTED]>

> Hello list,
>
>          I would really appreciate if someone could show me the correct
> way of using regexserde as i'm having some hard time using it. I have
> verified my regex through
> http://www.regexplanet.com/advanced/java/index.html and it's working fine
> there. But when i'm using the same pattern with regexserde i'm getting NULL.
>
> My input looks like this :
> <SOME_CHARACTER_STRING>
>
> and I want to extract the characters enclosed between the angle brackets.
>
> This is the command i'm using :
> hive> CREATE TABLE s(f1 STRING) ROW FORMAT SERDE
> 'org.apache.hadoop.hive.contrib.serde2.RegexSerDe'
>     > WITH SERDEPROPERTIES ( "input.regex" = "(\\w*)", "output.regex" > "%1$s") STORED AS TEXTFILE;
>
> LOAD works fine, but SELECT * gives me NULL.
>
> I am on hadoop-1.0.4 and hive-0.10.0
>
> Thank you so much for you time.
>
>
> Warm Regards,
> Tariq
> cloudfront.blogspot.com
>