[ https://issues.apache.org/jira/browse/NIFI-1249?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15042280#comment-15042280 ]
ASF subversion and git services commented on NIFI-1249: ------------------------------------------------------- Commit f378ee902127bd29b168c9bb15e991abe4eab0fa in nifi's branch refs/heads/master from [~markap14] [ https://git-wip-us.apache.org/repos/asf?p=nifi.git;h=f378ee9 ] NIFI-1249: Allow Processors to add their own variables to those referencable by Expression Language. Make ReplaceText allow users to reference back-references of regex matches > Allow ReplaceText expression language functions to access matching group > values > ------------------------------------------------------------------------------- > > Key: NIFI-1249 > URL: https://issues.apache.org/jira/browse/NIFI-1249 > Project: Apache NiFi > Issue Type: Improvement > Components: Extensions > Reporter: Joseph Witt > Assignee: Mark Payne > Fix For: 0.4.0 > > Attachments: > 0001-NIFI-1249-Allow-Processors-to-add-their-own-variable.patch, NIFI-1249.xml > > > Chanru from the user's mailing list asked how we'd support a pretty simple > use case of converting > {quote} > col1,col2,col3 > 2006,10-01-2004,10may2004 > 2007,15-05-2006,10jun2005 > 2009,8-8-2008,10aug2008 > {quote} > to this > {quote} > col1,col2,col3 > 2006,10-01-2004,2004-05-10 > 2007,15-05-2006,2005-06-10 > 2009,8-8-2008,2008-08-10 > {quote} > And we can do it today but it is surprising more effort than it should be. > The problem is that ReplaceText as-is gets us sooo close. But what it cannot > do is convert the original column three into a formatted Date object which > can then be written out as a formatted string. If it could then this would > be easy. So for example if we extracted each column as a matching group and > the replacement value could be > {code} > ${ $3:toDate('ddMMMyyy'):format('yyyy/MM/ddd')} > {code} > Then we'd be all set. But right now there is no way to take that third > matching group and do anything fun with expression language on it because it > is not a subject passed into the EL on which functions could be applied. If > instead we simply added those matching groups as keys/subjects available to > the EL then this becomes a quite powerful tool. -- This message was sent by Atlassian JIRA (v6.3.4#6332)