sort file removing duplicates possible?

Scott Sumner

ok, but why results are inconsistent (with large datasets)?

Various reasons, sometimes a regular expression approach to this needs to be refined to match the data better before it works well. If you search up some other threads on this topic you can trace through the evolution of a regex approach on certain datasets. However, I suspect you just want to get a workable solution and move on…and I fully endorse that. I’m tired of trying to use regex for this kind of thing. :)

BTW, @Claudia-Frank going the extra mile…installing UE trial version just to track this down…nice!

Claudia Frank

@Scott-Sumner

… boots are made for walking … :-)

I’m confused about how UE does sort and delete duplicates.
Used default settings.

Only sorting

Sorting and deleting duplicates

There is obviously something wrong about UEs algorithm, isn’t it?
And the version I used cut the list to 63732 lines.

Cheers
Claudia

patrickdrd

so… another number? amazing!

I’ve got version 16 which is lite somehow

Claudia Frank

me too - that is actually the latest one available for ubuntu.

Cheers
Claudia

patrickdrd

so the issue is: which result is the correct?

guy038

Hello, @patrickdrd, and All,

First of all, I quite pleased to be back on our N++ forum ! Indeed, I was away because of a general failure of my laptop hard drive C:, which, you could imagine, highly annoyed me and needed immediate care :-(((

So, after more than a week, a software purchase of EaseUS Recovery Wizard to restore my data ( the only one which could identify all my files, although Windows could not see the C: partition ), the fact that my first 32 Gb USB key did not work, either, ( Windows cannot format it !!), the re-install of the system on an other hard-disk, the Services Pack, the .NET versions, the different updates, some software installations and the total re-organization of my data structure, I can, now, close that bloody sub-routine !! ( Just note that, if I haven’t been able to get all my files back, I still had a last general backup, performed on the 04/13/18 )

Now, @patrickdrd, you said, in a post above :

guy038 regular expression results in 28109 lines

So, I download the list, from your link, in a new N++ tab :

https://easylist.to/easylist/easylist.txt

Notice that I, personally, found 69889 lines

Now, without any change on that text, I simply performed a sort with the N++ command Edit > Line Operations > Sort Lines Lexicographically Ascending, on that raw text

Then, I used the regex S/R, that I spoke in my previous post :

SEARCH : (?-s)(^.+\R)\1+

REPLACE : \1

I obtained a file of 69790 lines => The difference of 99 lines ( the duplicate ones ) were suppressed, almost immediately !

So, may be, I’m missing something ?

Cheers,

guy038

patrickdrd

I can’t say for sure, is there any tool that can tell us how many the unique lines are?

Claudia Frank

Ok - the difference of the total amount of lines in TextFX /python and regex
can be explained, the different usage of caseing(not sure if this is the right word).

Regex search/replace is insensitive whereas python script is sensitive.
Once we use the same, the result is the same.

For the provided example list, I assume, insensitive search and deletion is ok
but for other cases, sensitive search/replaces might be important.

Btw. I also used sort and uniq command line tools from linux with same results as python and regex.

Cheers
Claudia

patrickdrd

thanks for clarifying, in terms of speed?

can python script be made to ignore case?

I want to ignore case and I want it fast (useful for large files) too

patrickdrd

these are my settings:
https://i.imgur.com/AYmj5I4.jpg

and my version:
https://i.imgur.com/YRdLkSl.png

Scott Sumner

@patrickdrd

can python script be made to ignore case?

yes

I want to ignore case

Understood! :-D

and I want it fast (useful for large files) too

Ok, so what is “fast”? PS is in general going to be slower than some other methods, but unless you sort “large files”–also not well defined-- constantly why is speed all that important?

As an example, I took @Claudia-Frank 's PS one-liner above and ran it on the “easylist” data file, and for me it took 3.94 seconds…would that be defined as “fast” or “slow” or…?

And on the topic of “large” files, when you get too large (which honestly isn’t really big) you get into trouble with Notepad++ itself dealing with the files…

Claudia Frank

@patrickdrd

yes, with python we could do almost everything :-D

Is it important to keep the original lines untouched?
I mean, having two lines like

Test_line_content
test_Line_content

could result in either

test_line_content

or

Test_line_content

I will post a performance optimized script later today.

Cheers
Claudia

patrickdrd

@Scott-Sumner said:

@patrickdrd

can python script be made to ignore case?

yes

I want to ignore case

Understood! :-D

and I want it fast (useful for large files) too

Ok, so what is “fast”? PS is in general going to be slower than some other methods, but unless you sort “large files”–also not well defined-- constantly why is speed all that important?

As an example, I took @Claudia-Frank 's PS one-liner above and ran it on the “easylist” data file, and for me it took 3.94 seconds…would that be defined as “fast” or “slow” or…?

And on the topic of “large” files, when you get too large (which honestly isn’t really big) you get into trouble with Notepad++ itself dealing with the files…

3.94 for a 60k lines file is ok,
I mean I show much more time needed, if I remember correctly it should be with regular expressions

sorting is ok too, I want them sorted each time, if that’s what you ask scott

@claudia, if possible, make them “accent insensitive” too

Claudia Frank

@Patrick

sorry, don’t know the term “accent insensitive” , what does it mean?
For example that è is the same as e?

Can you provide example data (just need a couple of lines) to see if it is working correctly?
The speed test I will do with the easylist text.

Cheers
Claudia

Scott Sumner

I can envision the following specification for a general purpose script (goes beyond what @patrickdrd has asked for):

sort lines case sensitive, keep duplicate lines
sort lines case sensitive, remove duplicate lines
sort lines case sensitive, keep duplicate lines, reverse order on the sort
sort lines case sensitive, remove duplicate lines, reverse order on the sort
sort lines case insensitive, keep duplicate lines
sort lines case insensitive, remove duplicate lines
sort lines case insensitive, keep duplicate lines, reverse order on the sort
sort lines case insensitive, remove duplicate lines, reverse order on the sort
no sort, remove duplicate lines (case sensitive)
no sort, remove duplicate lines (case insensitive)

A selection active when invoking should define:

lines to be affected (only those touched by selection vertically); act on all lines if no selection
columns to be used as the sort key (if rectangular selection use the selected columns as the sort key; if zero-width rect selection consider the key to start at the selection column out to the end of the variable-length lines)

Probably I’ve forgotten something important to this “sort” of thing…

:-D

patrickdrd

@Claudia-Frank said:

@Patrick

sorry, don’t know the term “accent insensitive” , what does it mean?
For example that è is the same as e?

Can you provide example data (just need a couple of lines) to see if it is working correctly?
The speed test I will do with the easylist text.

Cheers
Claudia

yes, exactly that

Claudia Frank

OK - let’s see what we can do.

Cheers
Claudia

Scott Sumner

@Scott-Sumner

…the following spec…

Hey Scott! You did forget some things! How about when removing duplicates, we need the options to:

keep one occurrence of a duplicated line (when sorting)
keep no occurrences of a duplicated line (when sorting or not sorting)
keep LAST occurrence of a duplicated line (when not sorting)
keep FIRST occurrence of a duplicated line (when not sorting)

Claudia Frank

@Scott-Sumner

ahh - sorry too late - specs already defined for version 1 you need to open a feature request for version 2 :-D

Cheers
Claudia

Claudia Frank

You can find the 1st version of the script here.

In order to make it run there are two requirements, apart from the obvious one that you need to have python script plugin installed, which needs to be full-filled.

1.) be sue you have either installed the full package or download and unzip the TclTk into the NPP_INSTALL_DIR. Latest releases

2.) in order to make the “accent insensitive” feature working it is needed to install a python library called unidecode.
Unzip the .whl package into NPP_INSTALL_DIR\plugins\lib\

To check both requirements, open the python script console and do the following commands

 import Tkinter
 import unidecode

If you don’t see any errors - done.
Usage is simple - run the script and check the different options.

What should work is

sort/delete duplicates on whole text (aka nothing is selected)
sort/delete duplicates on vertically selected text

not supported yet:

sort/delete duplicates on rectangular selection

Cheers
Claudia

Btw. I spent most of the time creating this ugly window - so if someone wants to create a nicer gui - please go for it. I’m not really good in designing UIs.