Page 4 of 7
Re: A bit of regex help needed
Posted: 24 Mar 2014 16:47
by Marco
Re: A bit of regex help needed
Posted: 08 May 2014 12:07
by Andy Petrov
Hi, dear members!
Could someone please help me with regex?
Input:
FixedPart Variable1 Variable2.ext
Output:
FixedPart-Variable1Variable2.ext
Replace first space with - and trim second space.
TIA
Re: A bit of regex help needed
Posted: 08 May 2014 12:21
by highend
Re: A bit of regex help needed
Posted: 08 May 2014 14:18
by Andy Petrov
highend, thank you. It works
And how to make a more universal exp?
Input:
FixedPart Variable1.ext
FixedPart Variable1 Variable2.ext
FixedPart Variable1 Variable2 Variable3.ext
Output:
FixedPart-Variable1.ext
FixedPart-Variable1Variable2.ext
FixedPart-Variable1Variable2Variable3.ext
Replace first space with - and trim other spaces.
Re: A bit of regex help needed
Posted: 08 May 2014 14:46
by highend
In one turn? Sorry, don't know and no time to think about.
Just use a simple script:
Code: Select all
foreach($item, <get SelectedItemsNames |>, "|") {
renameitem(regexreplace(regexreplace($item, "(^.*?)\s", "$1-"), "\s"), $item, "3");
}
Reminder for myself: Coding is often faster than spending too much time on thinking about the problem...
All SELECTED files will be renamed according to your (last) wish.
Re: A bit of regex help needed
Posted: 08 May 2014 15:07
by Andy Petrov
Re: A bit of regex help needed
Posted: 17 May 2014 14:05
by nerdweed
From the tags.dat file, I am trying to read this
Read Filename and Extra Tag 4 for all those entries which has the tag Images
$data = RegExMatches ("$data", "(.+)\|0+\|.*Images.*\|.*\|.*\|(.*)\|.*\|", <crlf>) ;
The containers in this store this data however, RegExMatches returns the whole row.
$data = RegExReplace ("$data", "(.+)\|0+\|.*Images.*\|.*\|.*\|(.*)\|.*\|", "$1|$2") ;
If I try RegExRepalce, it returns the header data as well as well as files without Images tag.
Combination of both gives the desired output. Any better way to do this
Re: A bit of regex help needed
Posted: 17 May 2014 15:32
by highend
You shouldn't use needy matches only if really needed.
Apart from that regexmatches isn't able to capture specific parts via backreferences!
Imho you still have to combine them...
Code: Select all
text regexreplace(regexmatches(readfile("<xydata>\tag.dat"), "^[A-Z]:.*?\|0+\|.*?images.*?$", "<crlf>"), "^([A-Z]:.*?)\|0+\|.*?\|.*?\|.*?\|.*?\|(.*?)\|", "$1|$2");
Re: A bit of regex help needed
Posted: 17 May 2014 17:38
by nerdweed
Thanks highend
Didn't knew that ? is used for lazy method.
Re: A bit of regex help needed
Posted: 17 May 2014 17:46
by highend
Don't know what you mean.
Logical or is build with "|"
So a normal pattern looks like
"(one|two|three)"
Re: A bit of regex help needed
Posted: 05 Jun 2015 13:37
by SkyFrontier
Help on regexreplace, please.
These are samples:
AaaaaB - B must go
Aaaaa/AB2 - /AB2 must go
AaaaBCD/1AB2 - BCD/1AB2 must go
IOW, preserve everything before "/" BUT all the immediate left upper cased letters.
How?
Re: A bit of regex help needed
Posted: 05 Jun 2015 14:14
by highend
Code: Select all
$test = <<<>>>
AaaaaB
Aaaaa/AB2
AaaaBCD/1AB2
>>>;
$result = "";
$pattern = "^([A-Z][a-z]+)(/|[A-Z])(.*)$";
foreach($item, $test, "<crlf>") {
$result = $result . "Source: $item<crlf>Result: " . regexreplace($item, $pattern, "$1", 1) . "<crlf 2>";
}
text $result;
Re: A bit of regex help needed
Posted: 05 Jun 2015 14:46
by SkyFrontier
Hello, highend.
Thanks much!
Re: A bit of regex help needed
Posted: 06 Jun 2015 11:49
by SkyFrontier
Tried many variations but couldn't find a way to match the opposite way, this time with EXACT matches
/AAaaaa - aaaa
/AB2aaaa2 - aaaa2
/1ABaaaaBCD - aaaBCD
/AAB1aaaBCD - aaaBCD
-in this case, I'll have a line specifying that "AA", another line for "AB2", another for "/1ABa" (yes, a lower case there) and another for "/AAB1"- all these strings must go.
They appear like
/AAaaaa/AB2aaaa2
and will be output'ed like
aaaa
aaaa2
Thanks in advance.
Re: A bit of regex help needed
Posted: 06 Jun 2015 12:04
by highend
You can't catch
/1ABaaaaBCD - aaaBCD
with the same pattern you'd catch the other three ones... At least I don't see any logic behind it (no explicit number of characters to catch / delete).
Or does it follow the logic:
If a line begins with "/<a digit>" than capture the first 5 (including the "/") characters instead of only the first 4?
