That worked than you Joe
From: Joe Percivall [mailto:[email protected]]
Sent: Tuesday, December 15, 2015 10:57 AM
To: [email protected]
Subject: Re: Content replacement
Hey Chandu,
You messed up bit in the config of your ReplaceText processor. You should be
running "Line-by-line" mode so each line is treated separately.
Also your search value is a bit messed up. It was trying to use capture groups
in it ("$2"), it wasn't successfully handling an initial date of only 1 digit
("\d{2}"), it wasn't configured to match the spaces in column one and you need
to either need explicitly make groups "non-capture groups" using "?:" or don't
have them as capture groups. This search value works for me:
((?:\d{1,2}) (?:[a-zA-Z]{3}) (?:\d{4})),(.*),(.*)
Your replace value looks good though.
Hope that helps,
Joe
- - - - - -
Joseph Percivall
linkedin.com/in/Percivall
e: [email protected]<mailto:[email protected]>
On Monday, December 14, 2015 5:21 PM, Chandu Koripella
<[email protected]<mailto:[email protected]>> wrote:
Hi Mark,
I am testing replace test feature in 0.4.0. I don’t see it is working. I tried
with couple of the different columns. None of them aren’t replacing the text.
can you please take a look?
[cid:[email protected]]
Thanks,
Chandu
From: Mark Payne [mailto:[email protected]]
Sent: Friday, December 04, 2015 2:05 PM
To: [email protected]<mailto:[email protected]>
Subject: Re: Content replacement
Chandru,
This ticket has now been merged to master. This feature will be available in
0.4.0, which should be out next week. If you
prefer to build from source, it is now available for use from master. You
should be able to now configure ReplaceText to
easily replace the values:
[cid:[email protected]]
Here, we use the Search value of (.*?),(.*?),(\d+.*)
So the first Capturing Group gets the first column, the second Capturing Group
gets the second column, and the third Capturing Group gets the third column
(but only if it starts with a digit, so this won't match the header line and
the header line will remain unchanged).
For the Replacement Value, we use: $1,$2,${ '$3':toDate(
'ddMMMyyyy'):format('yyyy/MM/dd') }
So here we are using back references to replace the line with the first two
columns, followed by an Expression Language Expression that parses the third
back reference (the third column).
Since the variable that we want to reference is named $3, we need to enclose it
in quotes to escape the name because it begins with a non-alpha character. We
can then call any Expression Language
Expression that we want. So toDate() can be used to parse the string as Date
and then we can use format() to format that date as a string in a different
format.
You can also see the template available on the NIFI-1249 ticket, also attached
for convenience here, but I don't know if the apache mailing list will let the
template through.
This was a great addition to NiFi that I think will help out in a lot of cases
- thanks for reaching out to us on this!
Please let us know if you have issues getting this to work, or if you have any
further questions. We're happy to help however we can.
Thanks
-Mark
On Dec 4, 2015, at 3:44 PM, Joe Witt
<[email protected]<mailto:[email protected]>> wrote:
Chandru,
Correct what you're trying doesn't work just yet. But once this [1] is
reviewed and pushed to master we should have your original request covered
quite nicely.
[1] https://issues.apache.org/jira/browse/NIFI-1249
Thanks
Joe
On Fri, Dec 4, 2015 at 3:26 PM, Chandu Koripella
<[email protected]<mailto:[email protected]>> wrote:
Thanks Joe,
I believe I tried this option earlier. As you notice I have 4 groups in the
regular expression. I tried with all four groups. Nothing gets replaced in the
put file.
Group4 is 4 digit year, group3 is 3 digit month, group2 is date and group1 is
ddMMMyyyy
<image001.png>
From: Joe Witt [mailto:[email protected]<mailto:[email protected]>]
Sent: Thursday, December 03, 2015 7:00 PM
To: [email protected]<mailto:[email protected]>
Subject: Re: Content replacement
Your use case description was excellent and helped expose a powerful and simple
feature enhancement.
https://issues.apache.org/jira/browse/NIFI-1249
Thanks
Joe
On Thu, Dec 3, 2015 at 9:21 PM, Joe Witt
<[email protected]<mailto:[email protected]>> wrote:
Chandu,
In building a template to support this I'm finding it more involved than
desired. If I get that done will post it unless someone else does so sooner.
However, the only reason it is more involved than I'd like is that the
ReplaceText processor does almost everything we need here. It can be in line
mode to do the eval/replace on a per line basis and it can extract groups from
the matching line and use those groups to format a new output. However, the
difficulty in your example is that we cannot take that date string and convert
it to a date and then format it back to a string (right now). So will take a
look at how we could better support the use case as well.
Something else of note is that there are a couple of folks working on scripting
processors so you could break into a convenient script to help tackle some
cases like this as well.
Will dig a bit more and respond.
Thanks
Joe
On Thu, Dec 3, 2015 at 8:20 PM, Chandu Koripella
<[email protected]<mailto:[email protected]>> wrote:
HI,
I am very new to the nifi and evaluating to use to do simple transformations. I
am trying to convert date format from ddMMMyyyy to yyyy/mm/dd using replace
text option. I don’t find many options to do content replacement in the
documentation. it is more focused for attribute manipulation.
<image004.png>
Hsere is my simple file format. how do I transform the data in col3? I
appricaiate any help.
<image005.png>
Thanks,
Chandu
