You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hive.apache.org by "r7raul1984@163.com" <r7...@163.com> on 2015/02/25 07:36:51 UTC
Questions about Hive UDF cache under the hood
Hi everyone,
Consider the sql :
SELECT thumbnail( product_ image )
FROM advert i sements
WHERE product_name = ‘ Brownie’ ;
The product_ image field is a reference to a multi-megabyte image object. The thumbnail method reads in this object, and produces a small version of the image. Reading and processing the object can be quite time-consuming, and hence thumbnail can be very expensive. Since a given product may have many different advertisement layouts with the same photo, thumbnail may need to be computed many times on the same image object. It would be wasteful to actually invoke thumbnail on each reference to an image.
Question:
1 Hive support udf cache ?
2 How hive implement udf cache?
r7raul1984@163.com