question Hive2 vs Hive3 regexp_replace differences in Support Questions

Hive2 vs Hive3 regexp_replace differences

IanWilloughby — Thu, 03 Oct 2024 18:34:46 GMT

For a hive2 cluster, I do not have to escape a double quote inside a regexp_replace where the single quotes to start and end the regex.
But on hive3 in order for the code to work, the double quote needs to be escaped (Example to follow).
Why the difference?

-- Works in Hive 2 but not hive3 select id , case when url rlike '(?i)^https://www.linkedin.com/("?in|pub|company|profile)' then url else regexp_replace(url,'^https://www.linkedin.com/','https://www.linkedin.com/in/') end as social_url from db.table limit 10 -- Works in hive3 select id , case when url rlike '(?i)^https://www.linkedin.com/(\"?in|pub|company|profile)' then url else regexp_replace(url,'^https://www.linkedin.com/','https://www.linkedin.com/in/') end as social_url from db.table limit 10

Re: Hive2 vs Hive3 regexp_replace differences

DianaTorres — Thu, 03 Oct 2024 23:29:48 GMT

@IanWilloughby Welcome to the Cloudera Community!

To help you get the best possible solution, I have tagged our Hive experts @Shmoo @cravani who may be able to assist you further.

Please keep us updated on your post, and we hope you find a satisfactory solution to your query.

Re: Hive2 vs Hive3 regexp_replace differences

ggangadharan — Mon, 07 Oct 2024 08:02:30 GMT

@IanWilloughby Would it be possible for you to provide some sample records for the 'URL' column to help me gain a clearer understanding? Additionally, could you please share the specific versions of HIVE2 and HIVE3?

Re: Hive2 vs Hive3 regexp_replace differences

DianaTorres — Thu, 10 Oct 2024 23:28:31 GMT

@IanWilloughby If you are still experiencing the issue, can you provide the information @ggandharan has requested? Thanks.