topic Re: tFTPGet, Filemask and Regex question. in Talend Studio

tFTPGet, Filemask and Regex question.

Anonymous — Sat, 16 Nov 2024 13:52:41 GMT

Hello,
I?m trying to use Regex expression in Filemask of tFTPGet component.
I've never used regex before? So i use this site as a reference: http://www.regular-expressions.info/reference.html
How exactly regex expression should looks like if I want to get only files ending with _yyyyMMdd.csv
I tried to use the following expression but without success: ".*_.*\.csv"

Re: tFTPGet, Filemask and Regex question.

Anonymous — Wed, 15 Jul 2009 07:10:04 GMT

Hi,
For the purpose of unprofessional user, the file mask of tFTPXXX components is not according to the regex principle.
Try: "*_*.csv".
regard.

Re: tFTPGet, Filemask and Regex question.

Anonymous — Wed, 15 Jul 2009 14:50:25 GMT

Thanks for your answer nsun.
But if it?s not regex why the following regex style mask works fine - "bo(o|u)t" ?
By the way when I use ?windows style? mask *_??????.csv the following error appeals to java regex package java.util.regex.PatternSyntaxException.
Exception in component tFTPGet_1
java.util.regex.PatternSyntaxException: Dangling meta character '?' near index 5
.*_??????\.csv
^
What is correct syntax for Filemask in tFTPGet ?!

Re: tFTPGet, Filemask and Regex question.

Anonymous — Thu, 16 Jul 2009 07:26:57 GMT

==>But if it?s not regex why the following regex style mask works fine - "bo(o|u)t" ?
Because it actually use java.util.regex to handle matches in the bottom code, but before that , it do some transform, as a matter of fact: replace all "." with "\.") and replace all "*" with ".*".
==>By the way when I use ?windows style? mask *_??????.csv the following error appeals to java regex package java.util.regex.PatternSyntaxException.
I don't know the code writer's original intention, I guess it just in order to suit the behavior of common user who have no idea what Regex is, and it's not a completely implement of ?windows style? mask for sure. "*" is just easier to understand than ".*".
==>Exception in component tFTPGet_1
java.util.regex.PatternSyntaxException: Dangling meta character '?' near index 5
.*_??????\.csv
^
Above-mentioned, "??????" is kind of ?windows style? it doesn't support.
It indeed break the regex rule, so if needed we consider adding a feature of regex option for advanced settings.

Re: tFTPGet, Filemask and Regex question.

Anonymous — Fri, 17 Jul 2009 11:21:18 GMT

Thanks for your explanations nsun.
The following mask did a trick for my initial problem: ?*_{6}.csv?
You can use pretty much of regex in the filemask? just take in account that Taled replace ?.? with ?\.? and ?*? with ?.*?
I?m on TOS 3.2.0.M1 right now.