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

Switch to Threaded View
Pig, mail # user - Parallelism for small input data


Copy link to this message
-
Re: Parallelism for small input data
Dmitriy Ryaboy 2013-01-13, 22:54
"The udf (simple extends eval func) refers and reads a dictionary file of 6
MB for each input phrase."

Any reason to keep re-reading the dictionary instead of just reading it
once?

D

On Sun, Jan 13, 2013 at 4:47 AM, Dipesh Kumar Singh
<[EMAIL PROTECTED]>wrote:

> The udf (simple extends eval func) refers and reads a dictionary file of 6
> MB for each input phrase.
>