Hello, @abhay-gokhale, and All,

The is a safe way to sort with an offset, in Notepad++. The trick is to copy the key sort, in front of each line to be sorted ;-))

Let’s start with that sample text ( A long line from the license.txt file, split in several lines )

The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. This General Public License applies to most of the Free Software Foundation's software and to any other program whose authors commit to using it. (Some other Free Software Foundation software is covered by the GNU Library General Public License instead.) You can apply it to your programs, too.

I, simply, inserted a 5 digits number, surrounded by a space character, at column 11 of each line, giving the following text :

The licens 37529 es for most software are designed to take away your freedom to share and change 12345 it. By contrast, the GNU General Public License is intended to guarantee 94113 your freedom to share and change free software--to make sure the software i 88060 s free for all its users. This General Public License applies to most of th 46931 e Free Software Foundation's software and to any other program whose authors co 88060 mmit to using it. (Some other Free Software Foundation software is covered by 88060 the GNU Library General Public License instead.) You can apply it to your pr 62473 ograms, too.

That five digits number will be the expected key sort. Now, applying the regex S/R :

SEARCH (?-s)^.{11}(.{5}).+

REPLACE \1\x20\x20\x20$0

you should get the text :

37529 The licens 37529 es for most software are designed to take away your freedom to share 12345 and change 12345 it. By contrast, the GNU General Public License is intended to 94113 guarantee 94113 your freedom to share and change free software--to make sure the 88060 software i 88060 s free for all its users. This General Public License applies to 46931 most of th 46931 e Free Software Foundation's software and to any other program whose 88060 authors co 88060 mmit to using it. (Some other Free Software Foundation software is 88060 covered by 88060 the GNU Library General Public License instead.) You can apply it 62473 to your pr 62473 ograms, too.

Then, after running the sort operation ( Edit > Line Operations > Sort Lines Lexicographically Ascending, you get :

12345 and change 12345 it. By contrast, the GNU General Public License is intended to 37529 The licens 37529 es for most software are designed to take away your freedom to share 46931 most of th 46931 e Free Software Foundation's software and to any other program whose 62473 to your pr 62473 ograms, too. 88060 authors co 88060 mmit to using it. (Some other Free Software Foundation software is 88060 covered by 88060 the GNU Library General Public License instead.) You can apply it 88060 software i 88060 s free for all its users. This General Public License applies to 94113 guarantee 94113 your freedom to share and change free software--to make sure the

Notice that, when the key sort is identical ( case of 88060 ), for several lines, the lines are, in addition, sorted, according to the entire line contents ! ( authors - covered - software )

Finally, using this second simple regex S/R, below, we get rid of the temporary key sort, at column 1 :

SEARCH (?-s)^.{5}\h+

REPLACE Leave EMPTY

and obtain our expected text :

and change 12345 it. By contrast, the GNU General Public License is intended to The licens 37529 es for most software are designed to take away your freedom to share most of th 46931 e Free Software Foundation's software and to any other program whose to your pr 62473 ograms, too. authors co 88060 mmit to using it. (Some other Free Software Foundation software is covered by 88060 the GNU Library General Public License instead.) You can apply it software i 88060 s free for all its users. This General Public License applies to guarantee 94113 your freedom to share and change free software--to make sure the

Et voilà !

Best Regards,

guy038

Hello, @arpad-zsok, @scott-sumner and All,

Noticing that Arpad’s text contains, only, one | character per row, I suppose that this simple regex S/R, below, would be enough !

SEARCH [^"]+\|

REPLACE Leave EMPTY

But my solution isn’t the true goal of this post, indeed ! While running the Scott’s search regex, below :

(?-i)^((?:".+?",){11}")[0-9.]+?\|(?=NEGATIVE|POSITIVE)

against the text :

"processingIdentifier","dateTime","header/instrumentName","patientIdentifier","lastName","birthdate","sex","location","specimenIdentifier","testIdentifier","dilution","testName","result/value","unit","abnormalFlags","resultStatus","operatorName","resultDateTime","result","simple/resultCode" "A","B","C","D","E","20180328171007","VIDASPC01","F838-843","F838-843","SPT","UP Salmonella","0.07|NEGATIVE","F","20180328115445", "A","B","C","D","E","20180328171008","VIDASPC01","F838-844","F838-844","SPT","UP Salmonella","0.25|POSITIVE","F","20180328115446", "A","B","C","D","E","20180328171007","VIDASPC01","F838-843","F838-843","SPT","UP Salmonella","0.07|NEGATIVE","F","20180328115445",

I noticed that it took some time ( ~2s ), before finding the first expected zone :

“A”,“B”,“C”,“D”,“E”,“20180328171007”,“VIDASPC01”,“F838-843”,“F838-843”,“SPT”,“UP Salmonella”,"0.07|

And I tried to understand that behavior ! First, I realized that the matching delay was due to the headers line, which does not contain any | symbol. Thus, the problem could be simplified !

In a new tab, let’s build an unique line, with 3 initial “headers” lines, joined together, surrounded with two simple sentences :

This is a test "ProcessingIdentifier","dateTime","header/instrumentName","patientIdentifier","lastName","birthdate","sex","location","specimenIdentifier","testIdentifier","dilution","testName","result/value","unit","abnormalFlags","resultStatus","operatorName","resultDateTime","result","simple/resultCode","processingIdentifier","dateTime","header/instrumentName","patientIdentifier","lastName","birthdate","sex","location","specimenIdentifier","testIdentifier","dilution","testName","result/value","unit","abnormalFlags","resultStatus","operatorName","resultDateTime","result","simple/resultCode","ProcessingIdentifier","dateTime","header/instrumentName","patientIdentifier","lastName","birthdate","sex","location","specimenIdentifier","testIdentifier","dilution","testName","result/value","unit","abnormalFlags","resultStatus","operatorName","resultDateTime","result","simple/resultCode" This is a text

and let"s consider the generic regex ^(?-is)(".+?",){N}"z, where the letter N is any number > 0.Of course, as this text does not contain any lowercase letter z, the regex engine always answers Find: Can’t find the text…". On my old laptop, I obtained, depending of the used regex, the following times, for unmatching :

With regex ^(?-is)(".+?",){1}"z => Immediate answer

With regex ^(?-is)(".+?",){2}"z => Immediate answer

With regex ^(?-is)(".+?",){3}"z => 1s

With regex ^(?-is)(".+?",){4}"z => 2s

With regex ^(?-is)(".+?",){5}"z => 7s , with a wrong unique match ( all the file contents !)

Why the time, for the regex engine, to realize, that there is NO match, becomes exponential ? Well, this comes from the regex part ".+?". Indeed, to understand the process, let’s use the regex ^(?-is)(".+?",){1}"z, which can, also, be written ^(?-is)".+?","z

Actually, due to the lazy quantifier +?, this regex means : Find the smallest range of standard characters, between two " symbols, which is followed with the string ,"z !

So, let’s add a z lowercase letter, at the beginning of the second column header, as below :

"ProcessingIdentifier","zdateTime","header/instrumentName","patientIdentifier","lastName","birthdate",......

The regex part ".+?" catches the string “ProcessingIdentifier” and the part ,"z matches the ,"z string

Now, let’s add, instead, a z lowercase letter, at the beginning of the third column header, as below :

"ProcessingIdentifier","dateTime","zheader/instrumentName","patientIdentifier","lastName","birthdate",......

This time, the regex part ".+?" catches the string “ProcessingIdentifier”,“dateTime” and the part ,"z matches the ,"z string

And so on…, till inserting a z character at beginning of the last header :

"ProcessingIdentifier","dateTime","header/instrumentName",...................,"resultDateTime","result","zsimple/resultCode"

This time, the regex part ".+?" catches the huge string “ProcessingIdentifier”,“dateTime”…“resultStatus”,“operatorName”,“resultDateTime”,“result” ( 855 chars ) and the part ,"z matches the ,"z string, before the last header of the text ( simple/resultCode )

So , depending of the location of that z letter, the quantifier +? takes as many non-null characters as to meet the ,"z string ! This regex engine behavior can be considered as a foretracking operation ( by analogy to the backtracking one ! )

Now, everyone can understand why, if no z letter exists in the text, the time to answer becomes exponential, when replacing the {1} quantifier with greater values and, even with the {5} quantifier, I, personally, got a catastrophic foretracking :-((

Don’t forget that, for instance, the regex ^(?-is)(".+?",){11}"z can be rewritten :

^(?-is)".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?","z

And, as any part ".+?" may concern any form "....", "....","....", "....","....","....",…, it’s not difficult to guess that, because of the multiple possible combinations, troubles are not very far :-((

Two obvious solutions to that problem are to use the regexes ^(?-is)("[^,]+?",){11}"z or ^(?-is)("[^"]+?",){11}"z

However, I, also, got an other strange solution : to use the regex ^(?-is)(".+?",){11}+"z, with the possessive quantifier {11}+. Indeed, when no z letter exists in the list of headers, the negative answer, of the regex engine, is immediate !

At first sight, this {11}+ syntax seems quite weird, because the {11} quantifier cannot change and the backtracking/foretracking process does not seem to be involved with that quantifier ! But, actually, this means that the regex engine, will not try any combination, after a first unmatched search, in the regex part ".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",".+?",

And, if your add a z, at beginning of the 12th header ( “ztestName” ), you 'll get , with the regex ^(?-is)(".+?",){11}+"z, the only possible match :-))

In conclusion, the initial Scott’s regex should be used, with one the three syntaxes, below :

(?-i)^((?:"[^,]+?",){11}")[0-9.]+?\|(?=NEGATIVE|POSITIVE)

(?-i)^((?:"[^"]+?",){11}")[0-9.]+?\|(?=NEGATIVE|POSITIVE)

(?-i)^((?:".+?",){11}+")[0-9.]+?\|(?=NEGATIVE|POSITIVE)

Cheers,

guy038

I didn’t observe such issues with this version of N++. I’m using a similar configuration except of NppExec. Which version of NppExec do you use? 0.6 beta1?

I often think when it comes to computers and software, that if I come across a problem or in this case a “design behavior” as you call it, I’m not alone. :D
So I think of keeping the “Find” window open as an improvement and more users than me would find it useful.

Best regards

Get the answer,inspired and trying.

In the newest version, it’s not enough to delete all the .xml files in C:\Users\X\AppData\Roaming\Notepad++.
You must delete files in C:\Users\X\AppData\Roaming\Notepad++\backup as well.
Then just restart.

Very useful! Thank you!

Unfortunately, you’re not alone. Comptuer crashes can cause unexpected behavior, and so far, no one in the development team has been able to reliably replicate and thus debug how to prevent the problem. This is officially tracked in issue #2381

Regarding the specifics of your instance: you mentioned checking in %AppData%/Notepad++/plugins/config/NppFTP/Cache… but that’s only used to cache files if you’re using the NppFTP plugin to edit remote files (on your webserver, or similar). The right directory location for Notepad++'s backup (assuming you have Settings > Prefrences > Backup enabled) would be %AppData%\Notepad++\Backup: did you check there? Also, if you used a “local”/portable version (check ? menu > Debug Info, and look for “Local Conf mode:”; if it says ON, then it’s “local”), then the backup directory would be a subdirectory of your whatever directory notepad++.exe is in, rather than in the %AppData% hierarchy.

Unfortunately, if it’s not there, there’s probably not much we can do to help you recover it. However, you’re luckier than some, if your statement that “whatever data I had saved till yesterday is also fine” means that you did have your own backup somewhere, and that you’ve just lost the most recent edits. (If not, for the future, I highly recommend that you keep a backup of any important data**. This can be done with dedicated backup software, or with storing your files in a service like dropbox or one drive or google drive or similar. Or using version control software, such as git or subversion.)

History:

Back in December 2017, @chcg summarized the history of the topic up to that point: https://notepad-plus-plus.org/community/topic/14945/rare-null-write-glitch/2, which @Scott-Sumner updated in January (https://notepad-plus-plus.org/community/topic/15112/attempt-to-recover-files-after-npp-blank-2-files-after-error-missed-the-error-msg/2)

Since then, there have been at least a few more posts on the issue in this forum (incomplete):

https://notepad-plus-plus.org/community/topic/15549/help-my-pc-crashed-while-editing-and-deleted-my-whole-document/2 https://notepad-plus-plus.org/community/topic/15531/nul-nul-nul-problem-notpad

In March, I posted (https://notepad-plus-plus.org/community/topic/13302/fix-corrupted-txt-file-null/19) that I’d tried to force Windows to crash, but no matter what I did, I could not replicate the NUL-file issue. And @Scott-Sumner had tried the same thing some time earlier, as he revealed in the next post, with the same could-not-repeat-it results). This is a difficult issue to debug, and until someone steps forward with a reliable method of repeating it, it’s probably not going to get fixed.

@Herb-Martin

Many thanks Herb-Martin for your suggestion. Much appreciated.