[
https://issues.apache.org/jira/browse/NIFI-447?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14947884#comment-14947884
]
Mark Payne commented on NIFI-447:
---------------------------------
[~JPercivall] it looks like you mentioned Michael Payne in the last comment,
not me :)
Patch looks really good now. I did find one bug related to the character set.
There were two places where it was calling getBytes() without passing in the
appropriate character set. I went ahead and addressed those in a new patch. I
also fixed a couple of typos in the documentation and (at least I think)
clarified a bit of the documentation.
There were also a couple of performance improvements that I noticed could be
implemented. For example, since it's running over a few FlowFiles in each
iteration of onTrigger, I created a single byte array and reuse it rather than
creating a new byte array for each FlowFile. I also noticed that for the
Literal Replace, it was using a regular expression in order to insert the
Replacement Value before the line ending. This definitely appears to work okay
but can be pretty expensive, especially since the regex starts with a .* which
is a very expensive operation to perform in regex land. So I changed it a bit
to just copy the data character-by-character from the (buffered) input stream
to the (buffered) output stream until we find a \r or a \n and at that point
write out the replacement value, and then write out the rest of the characters.
I added unit tests for this to ensure that it's working right, and all of the
existing tests that you created pass as well.
I attached a new patch with these things merged in. Please review and make sure
that you're okay with the changes that I made.
Thanks
-Mark
> Add "Replacement Strategy" property to ReplaceText
> --------------------------------------------------
>
> Key: NIFI-447
> URL: https://issues.apache.org/jira/browse/NIFI-447
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Extensions
> Reporter: Mark Payne
> Fix For: 0.4.0
>
> Attachments:
> 0001-NIFI-447-Fixed-some-typos-and-clarified-some-documen.patch,
> NIFI-447.patch
>
>
> "Regular Expression" property should have displayName changed to "Search
> Value"
> ReplaceText should allow for a "Replacement Strategy" with values:
> Prepend (Search Value is ignored; replacement value prepended to FlowFile)
> Append (Search Value is ignored; replacement value appended to FlowFIle)
> Regex Replace (current behavior, so this should be default)
> Linewise Regex Replace (like current behavior if Evaluation Mode is set to
> Line-by-Line)
> Literal Replace (do not interpret the Replacement value
> The Line-by-Line property should be removed; this is breaking change!
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
