Searching random duplicate numbers/values in Notepad++

Mark Olson

@MariusGHub
If you had a smaller file, regular expressions would work for this. Because your file is so big, a regex-based solution would be abysmally slow (in technical terms, because duplicate-finding regexes scale as O(N^2), where N is the length of the file).

But it’s not hard to make a PythonScript script (follow link for help on installation, running) that would solve your issues.

# https://community.notepad-plus-plus.org/topic/24942/searching-random-duplicate-numbers-values-in-notepad
# what the script does: extracts an integer id from a record of the form <record from="Scandinavian/Scandinavian1363" to="graphics/pictures/person/2000650194/portrait"/>
# and eliminates every line that does not have the first instance of its id.
# EXAMPLE INPUT:
# <record from="blah1" to="graphics/pictures/person/2/portrait"/>
# <record from="blah2" to="graphics/pictures/person/1/portrait"/>
# <record from="blah3" to="graphics/pictures/person/1/portrait"/>
# <record from="blah4" to="graphics/pictures/person/3/portrait"/>
# <record from="blah5" to="graphics/pictures/person/7/portrait"/>
# <record from="blah6" to="graphics/pictures/person/3/portrait"/>
# OUTPUT FOR THAT INPUT:
# <record from="blah1" to="graphics/pictures/person/2/portrait"/>
# <record from="blah2" to="graphics/pictures/person/1/portrait"/>
# <record from="blah4" to="graphics/pictures/person/3/portrait"/>
# <record from="blah5" to="graphics/pictures/person/7/portrait"/>

def getFirstLineForEachDistinctId(text):
    lines = text.split('\r\n')
    first_line_distinct_id = []
    distinct_ids = set()
    for ii, line in enumerate(lines):
        # extract integer if it is followed by '"/portrait"/>' and then end of line
        id_ = re.findall('(\d+)/portrait"/>$', line)[0]
        if id_ not in distinct_ids:
            distinct_ids.add(id_)
            first_line_distinct_id.append(line)
    newtext = '\r\n'.join(first_line_distinct_id)
    editor.setText(newtext)
    
if __name__ == '__main__':
    getFirstLineForEachDistinctId(editor.getText())

I tested this on a file with 1 million lines and 200,000 distinct values and it executed in about a second.

MariusGHub

Thank you so much. I will take a look into it within half a hour cause I need to complete a race in an online league.
Curious how it works! I know phyton a bit from back in days when I was working at Honorbuddy profiles.

MariusGHub

I sadly get this error. I looked up the FAQ a couple of times, and retried it a couple of times too. I assume line 31 and 23 are the problem.
Edit: I did read on stackoverflow I’ve to ‘import re’?

    id_ = re.findall('(\d+)/portrait"/>$', line)[0]
NameError: global name 're' is not defined
Traceback (most recent call last):
  File "C:\Users\name\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\getFirstLineForEachDistinctId.py", line 31, in <module>
    getFirstLineForEachDistinctId(editor.getText())
  File "C:\Users\name\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\getFirstLineForEachDistinctId.py", line 23, in getFirstLineForEachDistinctId
    id_ = re.findall('(\d+)/portrait"/>$', line)[0]
NameError: global name 're' is not defined

Another one I got is after removing some text thats not the portrait line-stuff.

    id_ = re.findall('(\d+)/portrait"/>$', line)[0]
IndexError: list index out of range
Traceback (most recent call last):
  File "C:\Users\name\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\getFirstLineForEachDistinctId.py", line 31, in <module>
    getFirstLineForEachDistinctId(editor.getText())
  File "C:\Users\name\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\getFirstLineForEachDistinctId.py", line 23, in getFirstLineForEachDistinctId
    id_ = re.findall('(\d+)/portrait"/>$', line)[0]
IndexError: list index out of range

Looking up on the internet, hehe. So much to understand and it all feels like alien-language a bit. :D

Mark Olson

@MariusGHub
Sorry about that!
Here’s a revised version.

# https://community.notepad-plus-plus.org/topic/24942/searching-random-duplicate-numbers-values-in-notepad
# what the script does: extracts an integer id from a record of the form <record from="Scandinavian/Scandinavian1363" to="graphics/pictures/person/2000650194/portrait"/>
# and eliminates every line that does not have the first instance of its id.
# EXAMPLE INPUT:
# <record from="blah1" to="graphics/pictures/person/2/portrait"/>
# <record from="blah2" to="graphics/pictures/person/1/portrait"/>
# <record from="blah3" to="graphics/pictures/person/1/portrait"/>
# <record from="blah4" to="graphics/pictures/person/3/portrait"/>
# <record from="blah5" to="graphics/pictures/person/7/portrait"/>
# <record from="blah6" to="graphics/pictures/person/3/portrait"/>
# OUTPUT FOR THAT INPUT:
# <record from="blah1" to="graphics/pictures/person/2/portrait"/>
# <record from="blah2" to="graphics/pictures/person/1/portrait"/>
# <record from="blah4" to="graphics/pictures/person/3/portrait"/>
# <record from="blah5" to="graphics/pictures/person/7/portrait"/>
from Npp import *
import re

def getFirstLineForEachDistinctId(text):
    lines = text.split('\r\n')
    first_line_distinct_id = []
    distinct_ids = set()
    for ii, line in enumerate(lines):
        try:
            id_ = re.findall('(\d+)/portrait"/>$', line)[0]
            if id_ not in distinct_ids:
                distinct_ids.add(id_)
                first_line_distinct_id.append(line)
        except:
            first_line_distinct_id.append(line) # keep non-matching lines
    newtext = '\r\n'.join(first_line_distinct_id)
    editor.setText(newtext)
    
if __name__ == '__main__':
    getFirstLineForEachDistinctId(editor.getText())

MariusGHub

Thank you so very much. It worked within 1 second. I had 170 000 duplicates. I will immediately notify other users about my workaround and then your python-script. Will credit you ofcourse. I hope it will help lots of Football Manager users tidying up their Newgan config.ini!

Thank you again so much. Do you accept donations?

guy038

Hello, @mariusghub, @mark-olson and All,

Of course, Mark, in your script, we can change the line :

            id_ = re.findall('(\d+)/portrait"/>$', line)[0]

with this one, for the same results :

            id_ = re.findall('/(\d+)/', line)[0]

On the other hand, if you want to delete any duplicate line, this simple syntax seems to work :

            id_ = re.findall('.+', line)[0]

Now, I wanted to get some explanations on how to manage the python flags, in order to restrict the search, but I could not get any on-line help on the re.findall command or even findall. Could you enlight me, somehow, about it ? TIA !

Best Regards,

guy038

Alan Kilborn

@guy038 said in Searching random duplicate numbers/values in Notepad++:

but I could not get any on-line help on the re.findall command

Try: https://docs.python.org/3/library/re.html and then press Ctrl+f and search for findall. Or, since you are interested in the flags, maybe search for Flag constants are now instances.

But…I’m reluctant to provide this information as this really isn’t on-topic for this forum. Unless…you’re going to launch into some comparison of these flags versus some attribute of the Boost engine. Not sure where you’re going – not a mind reader – but hopefully you will be staying on-topic.

Actually…I’d say the script provided isn’t all that on-topic either. Ok, well, it DOES use a couple of editor functions, but really this is a thin veil (to read the entire text, and to write out a revised entire text) just to keep it a PythonScript. Really it is just a plain old Python program. It could just as well have been written in another language outside of Notepad+±- and I don’t think we want to get into writing specific-purpose non-Notepad++ related programs for people on this forum. If I were responding to the OP, I’d have said – short and sweet – “Notepad++ can’t help you with this; you’re going to have to do some programming”, and left it at that.

MariusGHub

Dear Alan,

I am very grateful Mark helped me with this issue. I was already trying for about nearly 6 hours with regex-commands. However, Mark pointed out my file was so big I had to use Python. I wouldn’t have a clue I could use Python and would been trying to use regex until I gave up.

I do believe also with the commands guy08 provided I learned that the code can be even shorter. And I also learned how this code works and what it does, and will likely help me in the future too. In the end I think this can benefit other future Notepad+±users with similar long files and want to lookup or delete duplicates.

For Football Manager I believe this tidying up even helped with the performance. Which is awesome of course!

It’s that we all live in this world and try to help each other. Give that push in the back and then start to learn marvelous things?

Kind regards,
Marius

guy038

Hello, @alan-kilborn, @mark-olson, @mariusghub and All,

Alan, thanks for your explanations. If I understand you correctly, this means that the @mark-olson’s Python text is a true Python program, whose all the stuff can still be interpreted as a Python Script, from within Notepad++ ?

This certainly explains why I could not find any documentation on the findall function, using only the Plugins > Python Script > Context-Help option !

Now, changing the line :

            id_ = re.findall('.+', line)[0]

by this one :

            id_ = re.findall('.+', line, re.I)[0]

on this text :

123
Test
test
TEST
Test
Test
test
Test
789

Still wrongly returns :


123
Test
test
TEST
789

Instead of the right result :

123
Test
789

Am I missing something obvious ?

BR

guy038

Alan Kilborn

@guy038 said in Searching random duplicate numbers/values in Notepad++:

this means that the @mark-olson’s Python text is a true Python program

No, not exactly. Mark’s script uses two editor object functions, which are specific to PythonScript and would generate errors if trying to run as a standalone Python program.

But as these editor object functions are merely grabbing all of the text from a Notepad++ tab, and setting all of the text from a Notepad++ tab later, they are standins for simple file reading/writing operations of standard Python.

I’m not going to harp on it too much, but it would be better to see PythonScript programs presented here focus on innovated tasks that PS can do for you in Notepad++.

Alan Kilborn

@guy038 said in Searching random duplicate numbers/values in Notepad++:

Am I missing something obvious ?

I’m being dragged kicking and screaming into off-topic land (but of course it is my choice to reply)… :-)

I could let Mark reply but this one is too easy:

The code in question is searching a single line for all matches, in a loop. Since one line doesn’t know about another, from one call of findall to the next, of course the output can contain duplicates (with and without regard to case).

PeterJones

Trying to bring it back to more PythonScript specifics:

lines = text.split('\r\n')
first_line_distinct_id = []
distinct_ids = set()
for ii, line in enumerate(lines):

If the function is just iterating through each line, why not use the editor.forEachLine(callback) syntax? It allows running a callback on each line in the file; the callback would then look at the line, and decide whether that line contains the first instance of an ID or not; you could then use editor.deleteLine() as shown in the example in the PythonScript documetation for forEachLine.

Or you could use the editor.research(regex, callback, ...), and have the regex find the full line containing the ID (and separate ID into group1); the callback can then track whether it’s found the particular ID yet or not: if it hasn’t found it, mark it as found and return the original string (so it keeps the first instance intact, because it replaces it with the same text); if it has been found, edit the string to remove that duplicate line (ie, return an empty string, so it replaces that line with an empty string)

Both of those implementation ideas would at least stay focused on features that are specific to PythonScript & Notepad++, rather than on generic python code.

Mark Olson

@PeterJones
I actually didn’t know about the editor.forEachLine callback. Had I known, my code could have been simplified.

I don’t particularly appreciate people (cf. AlanKilborn) nitpicking my solution, which was perfectly adequate for this problem.

But I figured I would also share another script I made for this general task, using an arbitrary regex instead of the specific one I made for this task:

# https://community.notepad-plus-plus.org/topic/24942/searching-random-duplicate-numbers-values-in-notepad/14?_=1695248025881
# finds all lines that match a certain regex (and have a subset of the line
# as capture group 1)
# and then:
# * removes lines that don't match the user-supplied regex
# * removes all lines except the first one for each value of the first capture group
# EXAMPLE INPUT:
# <record from="blah1" to="graphics/pictures/person/2/portrait"/>
# <record from="blah2" to="graphics/pictures/person/1/portrait"/>
# <record from="blah3" to="graphics/pictures/person/1/portrait"/>
# <record from="blah4" to="graphics/pictures/person/3/portrait"/>
# <record from="blah5" to="graphics/pictures/person/7/portrait"/>
# <record from="blah6" to="graphics/pictures/person/3/portrait"/>
# OUTPUT FOR THAT INPUT (WITH user-supplied regex "(\d+)/portrait"):
# <record from="blah1" to="graphics/pictures/person/2/portrait"/>
# <record from="blah2" to="graphics/pictures/person/1/portrait"/>
# <record from="blah4" to="graphics/pictures/person/3/portrait"/>
# <record from="blah5" to="graphics/pictures/person/7/portrait"/>
from Npp import *

__version__ = '0.0.1'

class REMDUPREGMATCH(object):
    def __init__(self):
        self.first_line_distinct_match = []
        self.distinct_matches = set()
        self.lines_matched = 0
        self.this_script_name = 'remove duplicate regex matches ' + __version__
        user_regex_with_group_1 = ''
        keep_non_matching_lines = True
        self.start_line_count = editor.getLineCount()
        def on_empty_line_match(mtch):
            self.start_line_count -= 1
        editor.research('^$', on_empty_line_match) # don't count empty lines
        eol = [ '\r\n', '\r', '\n' ][ editor.getEOLMode() ]
        while True:
            user_regex_with_group_1 = self.prompt('Enter regex with group 1 being the sort key.', user_regex_with_group_1)
            if user_regex_with_group_1 is None: return  # user cancel
            if not user_regex_with_group_1.strip():
                self.mb('Cannot specify an empty regex!  Try again.')
                continue
            if '(' not in user_regex_with_group_1:
                self.mb('Need to specify capture group 1 (as the sort key) in the regex!  Try again.')
                continue
            regex_err_msg = self.search_regex_is_invalid_error_msg(user_regex_with_group_1)
            if regex_err_msg:
                self.mb('Bad regular expression!\r\n\r\n{e}\r\n\r\nTry again.'.format(e=regex_err_msg))
                continue
            break
        # only get FIRST match on each line
        # also match BUT DON'T CAPTURE lines that don't match the regex
        # this allows us to preserve the non-matching lines
        user_regex_with_group_1 = '(?-s)^.*?%s.*$' % user_regex_with_group_1
        print('user_regex_with_group_1 =', user_regex_with_group_1)
        editor.research(user_regex_with_group_1, self.on_match)
        newtext = eol.join(self.first_line_distinct_match)
        editor.setText(newtext)
        if self.lines_matched != self.start_line_count:
            self.mb('Warning: %d matches were found but the document originally had %d non-empty lines' % (self.lines_matched, self.start_line_count))

    def on_match(self, mtch):
        self.lines_matched += 1
        # print(mtch.groups())
        matchval = mtch.group(1)
        if matchval not in self.distinct_matches:
            self.distinct_matches.add(matchval)
            self.first_line_distinct_match.append(mtch.group(0))

    def search_regex_is_invalid_error_msg(self, test_regex):
        try:
            # test test_regex for validity on a small subset of the document
            editor.research(test_regex, lambda _: None, 0, 0, 1000)
        except RuntimeError as r:
            return str(r)
        return ''

    def mb(self, msg, flags=0, title=''):  # a message-box function
        return notepad.messageBox(msg, title if title else self.this_script_name, flags)

    def prompt(self, prompt_text, default_text=''):
        if '\n' not in prompt_text: prompt_text = '\r\n' + prompt_text
        prompt_text += ':'
        return notepad.prompt(prompt_text, self.this_script_name, default_text)


if __name__ == '__main__':
    REMDUPREGMATCH()

Alan Kilborn

@Mark-Olson said in Searching random duplicate numbers/values in Notepad++:

I don’t particularly appreciate people (cf. AlanKilborn) nitpicking my solution, which was perfectly adequate for this problem.

Too bad? We try to keep things “on track” here. If something devolves into “I’ll write what is effectively not a Notepad++ solution” here, it shouldn’t be here.