[ https://issues.apache.org/jira/browse/PIG-2315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15311296#comment-15311296 ]
Daniel Dai commented on PIG-2315: --------------------------------- +1. Also note there is a performance regression in some cases. For example: {code} crawl = load 'webcrawl' as (url, pageid); extracted = foreach crawl generate flatten(REGEX_EXTRACT_ALL(url, '(http|https)://(.*?)/(.*)')) as (protocol:chararray, host:chararray, path:chararray); {code} Here the users just try to give additional information to Pig since REGEX_EXTRACT_ALL didn't declare types inside tuple and not intend to cast. With the change, Pig force a cast and there is no way to avoid that. The performance hit should be small and I believe it worth to clarify the syntax. > Make as clause work in generate > ------------------------------- > > Key: PIG-2315 > URL: https://issues.apache.org/jira/browse/PIG-2315 > Project: Pig > Issue Type: Bug > Reporter: Olga Natkovich > Assignee: Daniel Dai > Fix For: 0.17.0 > > Attachments: PIG-2315-1-rebase.patch, PIG-2315-1.patch, > PIG-2315-1.patch, pig-2315-2-after-rebase.patch, pig-2315-3-merged.patch > > > Currently, the following syntax is supported and ignored causing confusing > with users: > A1 = foreach A1 generate a as a:chararray ; > After this statement a just retains its previous type -- This message was sent by Atlassian JIRA (v6.3.4#6332)