Is it possible to use regex to achieve this?

Neil Schipper

Neil, test it against the text below

Do you really want a guy who was fairly certain that 548-488+1 = 81 to be the one to double check that which you’ve already tested?

Moking C

@guy038 thank you very much for trying to help. My data is not confidential so I will show you later (it’s on another computer in my office but today is my day-off).

Don’t know this changes anything but my data contains decimal point, like 0.12345. Terribly sorry for not being able to show the real data to you now, I will @ you again when i get back to work tomorrow. Thank you again.

Moking C

@neil-schipper thank you very much for you replies. Sorry that I am not familiar with that “code” function, I wanted to attach a file but couldn’t find an option. Like I said to guy038 I will show you the data when I get back to work tomorrow.

To answer your two questions, yes and yes. Although that “81” part confuses me a little, but I guess it would make things clearer for both of us when I show you the data later. For now, allow me to explain a little more:

the first line in every file is the starting date of the time-series data, then one line for one day for a 22 years. It starts from Jan 1st, 1976, and the missing data is from May 1st, 1977 to June 30, 1977. So I want to replace each day during these two months from the one of 1976 (it’s meteorological data, so this is one easy and reasonable solution). It looks like this:

19760101
0.1234
0.23456
0.12345
0.5678
…

Anyway, sorry that i couldn’t show you the real data now, I will @ you again with it tomorrow. Thank you for your help and patience.

Neil Schipper

Hi @moking-c,

Since the need is to copy and replace complete lines, the actual text on the lines is not relevant to the solution.
The 81 was an arithmetic error, which @guy038 detected and fixed.

So the next thing you should do is apply the regex in Guy’s last post, check the result and report back to us.

If there are problems and you want to provide text, paste it into a reply as described in the FAQ I linked. There is no file attach mechanism.

Moking C

@neil-schipper @guy038 Hi I’m back, I have tested with the solution guy038 proposed, and it worked as I wanted for line 488 to 548, which means every value from May-1-1977 to June-30-1977 is replaced by value of the same day from the previous year.

However there is a new problem, for some reason, 487 lines after the 548th line (namely from line 1036), another replacement happened for another consecutive 61-days, then another 487 lines after that the same thing, then again and agian until the end of the file. I tried with different files but the problem stays the same.

Here is the first 1650 lines of one file (8036 lines in total):

Neil Schipper

@moking-c

I could recreate your problem using Find expression:

(?:.*\R){122}\K((?:.+\R){61})((?:.+\R){304})(?:.+\R){61}

which is not the expression we suggested to you. I find that when using the expression we suggested:

(?-s)\A(?:.*\R){122}\K((?:.+\R){61})((?:.+\R){304})(?:.+\R){61}

the problem is not present.

Moking C

@neil-schipper hi, thank you for your reply. I was using the second expression all along but the problem kept happening. I wondered a possible reason and as it turned out: the problem disappeared after I updated my N++ to the newest version. Don’t know what possibly triggered the problem but I’m glad now it’s fine. So problem solved, thank you guys :)

Neil Schipper

@moking-c said:

the problem disappeared after I updated my N++

Good to hear. Do you know what the old version was? There have been changes over the years to npp’s regex component (based on seeing other discussions) but I’m not knowledgeable about which would pertain to this situation.

I wanted to mention that the data transformation this problem needed could have been solved (on a single file) without any F&R, regex or otherwise. A macro could have done the job. If you’ve never recorded and run macros, it’s a good skill to have. In this case, only normal cursor movement, text selection (shift-arrow) and copy/paste are needed, and, since goto line (ctl-g) is supported inside a macro, the solution is quite elegant. However, this approach would not have offered the ability to process many files in one super-process, so in this case it would have been far less desirable.

I also wish to draw your attention to the Donate link on the notepad-plus-plus.org main page. Presently, the link goes to a page with info on organizations working to provide assistance to distressed people in Ukraine.

Moking C

@neil-schipper sorry but I didn’t check the previous version before I updated it, so unless it’s possible to check it now I don’t know what it was, although I know I haven’t updated for quite a while, half a year at least, maybe even longer. And thanks for the suggestions, I will have a look!

Lycan Thrope

@moking-c
Well, you could do a search on your computer for files starting with the name npp or such, and see if it comes up. Simple Windows search may find the original install file.

guy038

Hi, @moking-c, @neil-schipper, @lycan-thrope and All,

The \A regex assertion, which anchors any further regex to the very beginning of a file, was improved and works properly ONLY SINCE the 7.9.1 N++ release !

Refer also at the end of this topic

Best Regards,

Neil Schipper

@guy038 said:

\A … works properly ONLY SINCE the 7.9.1 N++ release

Thanks, good to know. (I probably skimmed that factoid in the past but it wouldn’t have made an impression on me since I’m a novice user of \A.)

Refer also at the end of this topic

Those posts don’t say how \A behavior changed, and I’m not inclined to chase down and crawl through the Boost doc versions or developer issue discussions and commit comments to find out! If you happen to know, maybe share a sentence or two about how it behaved prior to 7.9.1, or link to something (brief!) where one can get a user-level idea of it.