Regex to find any lines that do NOT have a specific number of a character

Ekopalypse

@Mark-Yorkovich - sorry no answer just another question.

Why the hell is this not working?

Assuming the text

1|2|3|4|5|6|7|8|9|10
1|2|3|4|5|6|7|8|9
1|2|3|4|5|6|7|8|9|10|11

and using ^((.+?\|){9})(?!(.+?\|)) to find only lines with 10 columns.
It matches the line with 11 columns, why?

Mark Yorkovich

@Ekopalypse said:

^((.+?|){9})(?!(.+?|))

Yeah - you’re kinda highjacking my question. You’re looking for the complete opposite of what I am. Perhaps you could open your own post?

Just a suggestion…

Ekopalypse

@Mark-Yorkovich

Ja, I tried to solve it the other way. If you know the lines which do match the requirement
one could mark it, reverse the marking and now you have the lines which do not fulfill the
requirement.

Alan Kilborn

@Ekopalypse said:

reverse the marking

Eko means use bookmarking feature of the Mark command to match the lines you aren’t interested in finding, then invert the bookmarks (see the Search > Bookmarks menu) to get the lines you are interested in.

Alan Kilborn

This expression seems to put some redmarking (and thus you could bookmark on this basis) on only those lines that have exactly 9 pipes:

(?-s)^([^|\r\n]*?\|){9}(?!(?:.*?\|))

Using [^...] without putting line-ending characters inside makes me nervous, so I’ve done so above; not sure they are relevant here. :)

Obviously the {9} can be modified if nine is not really what is needed.

I am confused by why Eko’s attempt does not work.

Ekopalypse

@Alan-Kilborn

Alan, was this intentional?
(?-s)^([^|\r\n]*?\|){9}(?!(?:.*?\|))
or should it be
(?-s)^([^\|\r\n]*?\|){9}(?!(?:.*?\|))
(which by the way doesn’t seem to have any impact if used or not)

Now given your working example this works also
(?-s)^([^\|]*?\|){9}(?!.*?\|)

But I don’t understand why there is a need to make sure that a line
does not start with a pipe.

dinkumoil

@Ekopalypse

The following regex does the job: ^(?>.+?\|){9}(?!.+?\|). I’m not sure why but it seems to be related to the lack of backtracking due to ?> which turns group one to a non-capturing group.

@Mark-Yorkovich

In the Search & Replace dialog go to the Mark register:

Find what: ^(?>.+?\|){9}(?!.+?\|)
Bookmark line: ticked
Purge for each search: ticked
Wrap around: ticked
Regular expression: ticked

Click Mark All. Go to (menu) Search -> Bookmark -> Inverse Bookmark. Now all lines which do not contain exactly 9 pipe characters are bookmarked.

You can navigate to these lines with F2 (next bookmark) and SHIFT+F2 (previous bookmark).

You can also remove these lines by clicking (menu) Search -> Bookmark -> Remove bookmarked lines.

You can also do the opposite (removing not bookmarked lines) by clicking (menu) Search -> Bookmark -> Remove unmarked lines.

Mark Yorkovich

Eko’s exp works for me to find rows with 9 pipes/10 cols. Alan’s exp doesn’t match anything in my file with mostly 9 pipes/10 cols with a few known rows with less than 9 pipes.

I’m trying to match on rows with greater than or less than 9 pipes.

Alan Kilborn

@Ekopalypse

Was removing the escaping of the | inside the [ and ] intentional? Yes, I suppose, since it has no special meaning there and doesn’t need escaping.

don’t understand why there is a need to make sure that a line does not start with a pipe

I think that with this type of data, fields could be empty, thus if the first field is empty a line would start with a pipe? But, is the regex really saying what I think you implied? I’m saying “not pipe” not just at the start of a line, but for in between fields as well. And I’m only doing it this way because your original attempt using a . expression fails (for some odd and as yet unknown reason). I think I’m getting confused.

Ekopalypse

@dinkumoil

ok, I hope I finally understood this sentence

Match pattern independently of surrounding patterns, and don’t backtrack into it. Failure to match will cause the whole subject not to match.

which then means that my first attempt, which I was questioning, did backtrack.
which makes your regex is the one which I, and hopefully @Mark-Yorkovich were looking for.

@Alan-Kilborn,
Alan, ja, I guess you are right.

@Mark-Yorkovich, so does this work on your data and the procedure described by
@dinkumoil ?

dinkumoil

@Ekopalypse said:

ok, I hope I finally understood this sentence

I got the following hint at https://regex101.com/ when trying your regex:

A repeated capturing group will only capture the last iteration. Put a capturing group around the repeated group to capture all iterations or use a non-capturing group instead if you’re not interested in the data.

Then I read https://www.regular-expressions.info/atomic.html

Together it made me to give the non-capturing group a try.

Mark Yorkovich

@dinkumoil
I followed your instructions, but I’m not getting any matches.

Ekopalypse

@Mark-Yorkovich

make sure your caret is on the first line if you have not checked wrap around

Mark Yorkovich

@Ekopalypse
Yup, sure is. - No matches - double-checked my settings.

To reiterate: My file is mostly 9 pipes/10 cols per line, but some have less and a few more than that and I need to find those.

dinkumoil

@Mark-Yorkovich

I generated with the test data of @Ekopalypse a file of 146545 lines and did that what I’ve suggested above - I got the expected result.

Be sure that the pipe character in your file is really a pipe character (code 124). There is another one (code 166 in Windows-1252 character encoding) which looks nearly identical:

Pipe character: |
The other one: ¦

Mark Yorkovich

@dinkumoil said:

@Mark-Yorkovich

I generated with the test data of @Ekopalypse a file of 146545 lines and did that what I’ve suggested above - I got the expected result.

Be sure that the pipe character in your file is really a pipe character (code 124). There is another one (code 166 in Windows-1252 character encoding) which looks nearly identical:

Pipe character: |
The other one: ¦

Yup - they’re pipes.

Here is a good sample of what I’m working with. Lines 1, 9, 10, 11, 16 thru 20 and 36, 37 are single-line records with 9 pipes and 10 columns. Lines 2 thru 8 are one record and together have 9 pipes/10 cols. Similarly, lines 12 through 15 are a single record, and lines 21 thru 35 are a single record.

LOREM120|8 |3 |1 |1 |0 |0 |||INST020
LOREM120|9 |1 |1 |0 |0 |0 ||Lorem Ipsum Dolor]
LOREM: BS/BP

LOREM IPSUM:
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.|
IPSUM16|1 |1 |1 |1 |0 |0 |||3001479
IPSUM16|1 |2 |1 |1 |0 |0 |||3003077
IPSUM16|11 |0 |1 |0 |0 |0 |||
IPSUM16|13 |0 |1 |0 |0 |0 ||Lorem ipsum dolor sit amet
consectetur adipiscing elit,
sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.

DOLOR53	1	1	1	2		3003084
DOLOR53	2	3	1	1		Lorem ipsum
DOLOR53	2	4	1	1		Lorem ipsum
LOREM56	8	1	1	1		Lorem ipsum
LOREM56	8	2	1	1		Lorem ipsum
LOREM56	9	1	1	0	Lorem ipsum dolor sit amet

consectetur adipiscing elit

consectetur adipiscing elit
consectetur adipiscing elit

consectetur adipiscing elit
Lorem ipsum dolor sit amet
sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.

Lorem ipsum dolor sit amet
consectetur adipiscing elit
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.|
DOLOR19|1 |2 |1 |1 |0 |0 |||3003124
LOREM01|1 |1 |1 |1 |1 |0 |||3003024

Your suggested regex ^(?>.+?|){9}(?!.+?|) isn’t finding any matches on that

Ekopalypse

@Mark-Yorkovich

because it was assumed that all columns contain data

find: ^(?>.*?\|){9}(?!.*?\|) does not make that assumption.

Mark Yorkovich

@Ekopalypse said:

@Mark-Yorkovich
because it was assumed that all columns contain data

My bad. I didn’t give you all of the details of what I’m working with.

find: ^(?>.*?\|){9}(?!.*?\|) does not make that assumption.

This works.

So at this point what I’d need to do, ideally, is to do a Find/Replace, finding all of the new line/line feed characters - only in those now-bookmarked lines - and replace them with some other character (spaces, dummy chars, whatever) to get each of those records to be on one line. Can I do a find/replace on just the bookmarked lines? Or perhaps, instead of the multi-step approach, is there a way to do this on the Replace tab, entering a regex in the Find what box that finds those lines and just replace the new line characters with dummy characters in one step?

Alan Kilborn

@Mark-Yorkovich said:

Alan’s exp doesn’t match anything in my file

Well, if I copy and paste your “lorem ipsum” data (above) into a new tab and then run my regex (above) on it, I get lines with exactly 9 pipes redmarked, which I thought was the goal (or the inverse of the goal):

Imgur

So…I really don’t know where the disconnect is…

Alan Kilborn

@Mark-Yorkovich said:

…finding all of the new line/line feed characters - only in those now-bookmarked lines - and replace them with some other character (spaces, dummy chars, whatever) to get each of those records to be on one line

Didn’t we do all this the other day?