How to delete a complete tag with specific content

Scott Sumner · Jan 19, 2018, 1:26 PM

Unfortunately, Claudia’s regex can grab too much sometimes. For example, if run on the following data, it will select the first 3 lines with the first Find Next press (good!), but the second Find Next press will match the 6 lines after that (BAD! as that contains a non-zero “Active” field and a replace with “nothing” operation would delete that!):

<Report>
    <Active>0000000000</Active>
</Report>
<Report>
    <Active>0000000001</Active>
</Report>
<Report>
    <Active>0000000000</Active>
</Report>

Try this instead:

Pre-step: Back up your data!
Find what zone: (?s-i)<Report>.*?<Active>(?:(0000000000)|(\d+))</Active>.*?</Report>\R
Replace with zone: (?1:$0\r\n
Search mode: Regular expression
Wrap around checkbox: Ticked
Action: Press the Replace All button

Christoph Kahle · Jan 19, 2018, 1:33 PM

Still the same.
Now a lager sample

<Report>
	<UserName>Angebot </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000338</IndexPerm><TempIndex>0000000338</TempIndex></Report>
<Report>
	<UserName>Angebot.</UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000339</IndexPerm><TempIndex>0000000339</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000340</IndexPerm><TempIndex>0000000340</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000341</IndexPerm><TempIndex>0000000341</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000342</IndexPerm><TempIndex>0000000342</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000343</IndexPerm><TempIndex>0000000343</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000344</IndexPerm><TempIndex>0000000344</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000345</IndexPerm><TempIndex>0000000345</TempIndex></Report>
<Report>
	<UserName>Artikel mit hinterlegten Freifeldern </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>Artikelliste</Filter>
	<VMB>artlist12.vmb</VMB>
	<Active>0000000001</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<UserName filter="handwerk hw_plus">Material mit hinterlegten Freifeldern 1-6</UserName><IndexPerm>0000000346</IndexPerm><TempIndex>0000000346</TempIndex></Report>
<Report>
	<UserName>Kunden mit hinterlegten Freifeldern </UserName>
	<UserName filter="Lbb Stb">UStIdNr-Anfragen Mandanten</UserName>
	<RefReport>Drucker</RefReport>
	<Filter>Kundenliste</Filter>
	<VMB>KundFrei1.vmb</VMB>
	<ChangeTarget>Nein</ChangeTarget>
	<Remark/>
	<Argument1/>
	<Argument2/>
	<UseProgram>Ja</UseProgram>
	<Active>0000000001</Active>
	<Group>Kundenliste</Group>
	<CommandID>0x800B</CommandID>
	<PreviewBmp>KunUstID.png</PreviewBmp>
	<IndexPerm>0000000347</IndexPerm><TempIndex>0000000347</TempIndex></Report>
<Report>
	<UserName>Lieferanten mit hinterlegten Freifeldern 1-6</UserName>
	<RefReport>Drucker</RefReport>
	<Filter>Lieferantenliste</Filter>
	<VMB>liefFrei1.vmb</VMB>
	<Remark></Remark>
	<Argument1/>
	<Argument2/>
	<UseProgram>Ja</UseProgram>
	<Active>0000000001</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<Group>Lieferantenliste</Group>
	<CommandID>0x8038</CommandID>
	<PreviewBmp>KunUstID.png</PreviewBmp>
	<IndexPerm>0000000348</IndexPerm><TempIndex>0000000348</TempIndex></Report>
<Report>
	<UserName>Angebot </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000349</IndexPerm><TempIndex>0000000349</TempIndex></Report>
<Report>
	<UserName>Rechnung </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro.vmb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000001</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000350</IndexPerm><TempIndex>0000000350</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000351</IndexPerm><TempIndex>0000000351</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000352</IndexPerm><TempIndex>0000000352</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000353</IndexPerm><TempIndex>0000000353</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000354</IndexPerm><TempIndex>0000000354</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000355</IndexPerm><TempIndex>0000000355</TempIndex></Report>
<Report>
	<UserName>Auftragsbestätigunglogo </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro.vmb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000001</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000388</IndexPerm><TempIndex>0000000388</TempIndex><Group></Group><Remark></Remark><Argument2></Argument2><UseProgram>Ja</UseProgram></Report>
<Report>
	<UserName>Gutschrift </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro.vmb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000358</IndexPerm><TempIndex>0000000358</TempIndex><Group></Group><Remark></Remark><Argument2></Argument2><UseProgram>Ja</UseProgram></Report>
<Report>
	<UserName>Gutschrift </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro.vmb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000001</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000357</IndexPerm><TempIndex>0000000357</TempIndex><Group></Group><Remark></Remark><Argument2></Argument2><UseProgram>Ja</UseProgram></Report>
<Report>
	<UserName>Gutschrift old</UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro.vmb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000356</IndexPerm><TempIndex>0000000356</TempIndex><Group></Group><Remark></Remark><Argument2></Argument2><UseProgram>Ja</UseProgram></Report>
<Report>
	<UserName>Gutschrift </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro.vmb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000360</IndexPerm><TempIndex>0000000360</TempIndex><Group></Group><Remark></Remark><Argument2></Argument2><UseProgram>Ja</UseProgram></Report>
<Report>
	<UserName>eigene Bestellungen</UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000001</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000359</IndexPerm><TempIndex>0000000359</TempIndex></Report>
<Report>
	<UserName>Angebot  </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000361</IndexPerm><TempIndex>0000000361</TempIndex></Report>
<Report>
	<UserName>Angebot </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000362</IndexPerm><TempIndex>0000000362</TempIndex></Report>
<Report>
	<UserName>Angebot </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000363</IndexPerm><TempIndex>0000000363</TempIndex></Report>
<Report>
	<UserName>Angebot </UserName>
	<RefReport>Drucker</RefReport>
	<Filter>netto Auftrag brutto layout</Filter>
	<VMB>lay_pro_gue.umb</VMB>
	<Argument1>layout.ini</Argument1>
	<Active>0000000000</Active>
	<ChangeTarget>Nein</ChangeTarget>
	<IndexPerm>0000000364</IndexPerm><TempIndex>0000000364</TempIndex></Report>

Scott Sumner · Jan 19, 2018, 1:38 PM

@Christoph-Kahle

Well, trying my solution on that new data works for me!

Claudia Frank · Jan 19, 2018, 1:47 PM

I see my match get’s expanded until a report tag happens with action 000000.
Scotts version is working for me as well.

Cheers
Claudia

Claudia Frank · Jan 19, 2018, 2:07 PM

@Christoph-Kahle , @Scott-Sumner

find what:

(?s)<Report>(?:(?!</Report>).)*<Active>0000000000</Active>.*?</Report>

replace with empty

This seems to work as well :-D

Cheers
Claudia

Scott Sumner · Jan 19, 2018, 2:07 PM

So some explanation is probably in order.

The Find what will match EVERY <Report> thru </Report> (with line-ending after the closing tag) one-by-one as it moves thru the file. The difference is that when an all-zeroes Active field is encountered, it gets saved into “group #1”.

The Replace with is where the difference comes in: If “group #1” was matched during the Find, the replacement text is “nothing” (thus a delete operation), otherwise it is the entire match (represented by $0) from the find stage.

The ?1 is a “test” at replacement time. It’s general form is ?1a:b and means "if group #1 was matched at find time, replace with “a” at replace time, otherwise replace with “b”. In our specific case above “a” is omitted, which means “insert nothing here”.

guy038 · Jan 21, 2018, 1:45 AM

Hello, @christoph-kahle, @claudia-frank, @scott-sumner and All,

Now, most members of N++ community, who use regexes, in search/replacement operations, know the main difference between a lazy and a greedy quantifier ! For instance, given the * quantifier, which is a shortcut of the syntax {0,} :

The (?-s)0.*9 regex matches the greatest range of characters, between digits 0 and 9 of a same line
The (?-s)0.*?9 regex matches the tallest range of characters, between digits 0 and 9 of a same line
The (?s)0.*9 regex matches the greatest range of characters, between digits 0 and 9, even of several lines
The (?s)0.*?9 regex matches the tallest range of characters, between digits 0 and 9, even of several lines

So, assuming the one-line text, below :

1234567890<Report>First Block</Report>123457890<Report>Second Block</Report>1234567890<Report>Third Block</Report>1234567890

You should, easily, see the difference between the regex (?-s)<Report>.*</Report> and the regex (?-s)<Report>.*?</Report>

Just note that the later regex could be rewritten, as well, (?-s)<Report>((?!</Report>).)*</Report> . Indeed, this regex means :

Find a zone, beginning with <Report>, ending with </Report> and which does not contain, at any position, after <Report>, till </Report>, the string </Report> !

Now, why the first form the the Claudia’s regex, below, is not totally correct ?

(?s)<Report>.*?<Active>0000000000</Active>.*?</Report>

I, probably, would have build this one, at first sight, too :-) Indeed, this regex looks for the nearest string <Active>0000000000</Active>, after the start tag <Report>, itself followed by the nearest end tag </Report>. Everything seems OK…

However, we do a mistake because the fact of reaching the nearest <Active>0000000000</Active> and the fact of reaching the nearest </Report> are independent events ! And we may find, first, a <Active>0000000000</Active> block, after crossing many end tags </Reports :-((

So, in order to tell the regex engine to find, first, the nearest string <Active>0000000000</Active>, of the SAME block <Report>.....<Report>, we must use the second Claudia’s regex, below :

(?s)<Report>(?:(?!</Report>).)*<Active>0000000000</Active>.*?</Report>

Notes :

No need to change the first * greedy quantifier, before <Active>, in its lazy form ! Actually, due to the negative look-ahead, you already know that the range, between <Report> and <Active>..., is part of the current <Report>...</Report> zone, which always contains, in Christoph’s file, an unique block <Active>.....</Active> !
On the contrary, the last *? lazy modifier, before </Report>, is, of course, mandatory, to get the end of the current <Report> tag

Cheers,

guy038

Scott Sumner · Jan 22, 2018, 1:50 AM

@guy038 said:

Find a zone, beginning with <Report>, ending with </Report> and which does not contain, at any position, after <Report>, till </Report>, the string </Report> !

Hi Guy! I’m sure you can appreciate the humor in your statement. Of course the string </Report> isn’t going to occur before the string </Report> ! Because if it did, then it would have! ;)

guy038 · Jan 22, 2018, 4:07 AM

Hi, @scott-sumner,

Yes, my statement looks a bit weird ! But, English is not my mother language, and probably, a better formulation exists, in fluent English-American :-)

Oh !, Perhaps, I should had write :

Find a zone, beginning with <Report>, ending with </Report> and which does not contain, at any position, after <Report>, till </Report>, AN OTHER string </Report> ?

Actually, I just wanted to point out, that the simple case (?-s)a.*?b, with a lazy quantifier, may, as well, be written, with a greedy quantifier :

(?-s)a((?!b).)*b
a[^b\r\n]*b

Cheers,

guy038

PeterJones · Jan 22, 2018, 2:08 PM

@Scott-Sumner,

You’ve got to admit, even for native speakers, it’s sometimes hard to translate a regular expression into clear English. I would probably phrase the non-greedy version as

Find a zone, beginning with <Report> and ending with the first instance of </Report>

and the greedy version as

Find a zone, beginning with <Report> and ending with the last instance of </Report>

Luiz Antonio Souza Ribeiro · Jun 13, 2020, 10:11 PM

@guy038 said in How to delete a complete tag with specific content:

(?s)<Report>(?:(?!</Report>).)<Active>0000000000</Active>.?</Report>

Thank you. It works for me in a similar case.