search and replace / regEx

Frank Kirschner

Hi community.
I have a corrupted iCalendar file and want to delete with search and replace the waste between two tags:

END:VCALENDAR
3F68636A-A88D-4B6D-95C7-DC5B65910335.ics ZÔúƒ4504b552cef6ac7c1141ef12fba9a94a ²VEVENT ZÚ p ZÚˆ3F68636A-A88D-4B6D-95C7-DC5B65910335€ $ (È„ P€ „ Í`õ ¢BEGIN:VCALENDAR

Everything between END:VCALENDAR and BEGIN:VCALENDAR should be delete.
How do I have to use the search pattern?

Thanks, and best regards,
Frank

Claudia Frank

@Frank-Kirschner

using regular expression in find dialog search mode,
find what is

(?s)(?<=BEGIN:VCALENDAR).*?(?=\REND:VCALENDAR)

and replace with stays empty, then press Replace All.
Note, you cannot use Replace to jump through (bug) the file
and see what gets replaced.

Cheers
Claudia

Scott Sumner

@Claudia-Frank

Did you get BEGIN and END mixed up? I found that your regexp didn’t work on the sample data, but this one seems to (the biggest change is to swap BEGIN and END):

(?s-i)(?<=END:VCALENDAR).*?(?=BEGIN:VCALENDAR)

Perhaps the OP would like to be able to individually cycle thru the matches. The above solutions don’t allow that but this one does:

Find what zone: (?s-i)END:VCALENDAR.*?BEGIN:VCALENDAR
Replace with zone: END:VCALENDARBEGIN:VCALENDAR
Search mode: Regular expression

Claudia Frank

@Scott-Sumner

Hi Scott,
thanks for the head up - yes, I did - I was just thinking BEGIN comes before END but I guess that is where OPs issue comes from.

But the regex itself, when switching the end/begin terms should work, shouldn’t?

(?s)(?<=END:VCALENDAR).*?(?=\RBEGIN:VCALENDAR)

does it for me.

Cheers
Claudia

Scott Sumner

@Claudia-Frank

Starting with your original regexp, I added the -i because the OP’s spec was definitely uppercase and I deleted your \R because there didn’t seem to be a requirement that a line-ending occurred before the BEGIN, and indeed in the sample data there doesn’t appear to be one? Thus, copying the OP’s sample data to a N++ tab and trying your original regexp on it yielded no matches for me…hmmm…

Claudia Frank

@Scott-Sumner

You are right, when copying the sample data there seems to be no eol but from what is displayed I assumed there is one
but even without \R it seems to work for me

This cannot be an linux/windows issue, can it be?

Note, I shorten the line to fit into the screen - but it works with the original data as well.

Cheers
Claudia

Claudia Frank

Ah, now I see, I guess - you used the original data without the eol but with my regex which included the eol.
Yes, makes sense, does not work.

Cheers
Claudia

guy038

Hello, @frank-kirschner, @claudia-frank, @scott-sumner and All,

I thought about a third regex which, in addition, looks if :

The END:VCALENDAR string is preceded by a line-break
The BEGIN:VCALENDAR string is followed by a line-break

and, in replacement, this regex S/R adds a line-break, if not initially present, in O.P.'s text

So, assuming the four possible cases, below :

blah blah
bla bla blaEND:VCALENDAR
3F68636A-A88D-4B6D-95C7-DC5B65910335.ics ZÔúƒ4504b552cef6ac7c1141ef12fba9a94a ²VEVENT ZÚ p ZÚˆ3F68636A-A88D-4B6D-95C7-DC5B65910335€ $ (È„ P€ „ Í`õ ¢BEGIN:VCALENDARblah blah blah
bla bla...


blah blah
bla bla blaEND:VCALENDAR
3F68636A-A88D-4B6D-95C7-DC5B65910335.ics ZÔúƒ4504b552cef6ac7c1141ef12fba9a94a ²VEVENT ZÚ p ZÚˆ3F68636A-A88D-4B6D-95C7-DC5B65910335€ $ (È„ P€ „ Í`õ ¢BEGIN:VCALENDAR
blah blah blah...
bla bla...

blah blah
bla bla bla
END:VCALENDAR
3F68636A-A88D-4B6D-95C7-DC5B65910335.ics ZÔúƒ4504b552cef6ac7c1141ef12fba9a94a ²VEVENT ZÚ p ZÚˆ3F68636A-A88D-4B6D-95C7-DC5B65910335€ $ (È„ P€ „ Í`õ ¢BEGIN:VCALENDARblah blah blah
bla bla...


blah blah
bla bla bla
END:VCALENDAR
3F68636A-A88D-4B6D-95C7-DC5B65910335.ics ZÔúƒ4504b552cef6ac7c1141ef12fba9a94a ²VEVENT ZÚ p ZÚˆ3F68636A-A88D-4B6D-95C7-DC5B65910335€ $ (È„ P€ „ Í`õ ¢BEGIN:VCALENDAR
blah blah blah
bla bla...

then the regex S/R :

SEARCH (?s-i)((\R)?END:VCALENDAR).*?(BEGIN:VCALENDAR(\R)?)

REPLACE (?2:\r\n)\1\r\n\3(?4:\r\n)

would gives the following text ( four identical blocks of text ) :

blah blah
bla bla bla
END:VCALENDAR
BEGIN:VCALENDAR
blah blah blah
bla bla...


blah blah
bla bla bla
END:VCALENDAR
BEGIN:VCALENDAR
blah blah blah...
bla bla...

blah blah
bla bla bla
END:VCALENDAR
BEGIN:VCALENDAR
blah blah blah
bla bla...


blah blah
bla bla bla
END:VCALENDAR
BEGIN:VCALENDAR
blah blah blah
bla bla...

Et voilà !

Notes :

You may, either, click several times on the Replace button or once, only, on the Replace All button
In search :
- First the (?s-i) modifiers forces :
  - The search to be performed in a sensitive way ( NON-insensitive ! )
  - The special dot character . to be considered as any single character, even an End of Line one
- Then group 1 contains the string END:VCALENDAR, possibly preceded with a line-break
- The part (\R)? ( identical to the form (\R){0,1} ) represents an optional line-break ( group 2 )
- Now, the .*? part ( identical to .{0,}? ) stands for the smallest range of any character, between the two strings END:VCALENDAR and BEGIN:VCALENDAR
- The group 3 contains the string BEGIN:VCALENDAR, possibly followed with a line-break
- Finally, the part (\R)? ( identical to the form (\R){0,1} ) represents an optional line-break ( group 4 )
In replacement :
- The conditional replacement feature (?2:\r\n), rewrites a line-break, only if group 2 ( \R ) does not exist, before the string END:VCALENDAR
- The block \2\r\n\3 adds the strings END:VCALENDAR and BEGIN:VCALENDAR, separated with a line-break ( \r\n )
- The conditional replacement feature (?4:\r\n), rewrites a line-break, only if group 4 ( \R ) does not exist, after the string BEGIN:VCALENDAR

Cheers,

guy038