Add line of text to beginning of multiple files

Kevin Hemken

Thanks. Works perfect!

guy038

Hello Kevin, Claudia and All,

An other possibility would be :

SEARCH (?s).*

REPLACE WHATEVERTEXT\r\n$0

NOTES :

The (?s) modifier forces the regex engine to consider that the meta character dot . matches, absolutely, any single character, whatever the . matches new line option is set or unset
The $0 form , in replacement, stands for the entire searched regex, that is to say, the actual contents of each scanned file

REMARK :

Due to a N++ regex engine bug, about backward assertions, we cannot use, for instance, the \A assertion, which, normally, represents the very beginning of a file. IF it would have been possible, the S/R would have been, simply :

SEARCH : \A

REPLACE : WHATEVERTEXT\r\n

Therefore, the Claudia’s work-around, is a nice solution to that bug !!

Best Regards,

guy038

Claudia Frank

Hi @guy038,

nice one ( or two :-)) but where did you find the ?s ?
I can’t seem to see it in the official docs.

Cheers
Claudia

Robin Cruise

Search: ^\A(.*)$
Replace by: ANYTHING \r$1

or

Search: ((?s:.*))
Replace by: ANYTHING \n\1

aatmankothari

@Claudia-Frank Thanks a lot!
This saved me a lot of time as I had around 500 files to be modified.

Brad Stewart

@Robin-Cruise I tried your first option an it worked perfectly.

guy038

Hi, @Kevin-Hemken, @Robin-Cruise, @aatmankothari, @Brad-Stewart, @Terry-r, @scott-sumner, @claudia-frank and All,

As explained in the post, below :

https://community.notepad-plus-plus.org/post/33823

The \A assertion which should match the very beginning of current file does not work properly, whereas the opposite assertion \z does match the very end of current file, in all cases !

Take the time to read all the discussion and, especially, the @scott-sumner’s opinion on the \A problem and, more generally, on all the backward assertions ! Finally, I think that his approach is the right one ;-))

https://community.notepad-plus-plus.org/post/34704

In this discussion, @Terry-R did found out a nice work-around to replace the bugged \A feature with the regex (?<!\x0A)^ and I proposed the more rigorous regex (?<!\n|\r|\f)^, because the dot regex does not match the FF control character !

However, thanks to the @Brad-Stewart’s reply, I saw the @robin-cruise’s reply, too and, particularly, his first regex S/R: I was intrigued with the beginning ^\A. After some tests, I can say that this regex is the shortest one, which works properly, whatever the current location of the caret, as long as you’re using the global replacement, with a click on the Replace All button, exclusively and that the Wrap around option is ticked ;-))

So, in order to add the expression ANY TEXT THAT YOU WANT TO INSERT AT THE VERY BEGINNING at the very beginning of the current file or all opened documents or all the files from a directory, simply use the following regex S/R :

SEARCH ^\A

REPLACE ANY TEXT THAT YOU WANT TO INSERT AT THE VERY BEGINNING

You may insert an entire line, with the syntax :

SEARCH ^\A

REPLACE ANY TEXT THAT YOU WANT TO INSERT AT THE VERY BEGINNING\r\n ( Remove \r in case of an Unix file )

Remark :

In case of your replacement regex ends with the forms \r\n, \n or \r, depending of your file(s) structure, as above, it/they must not begin with one/several true empty lines !

Best Regards,

guy038

P.S. :

Assuming a file, containing a lot of lines ============, which act as separators, with some text in between, then the regex which matches the 6th separator line, only, could be, with the @robin-cruise’s solution :

SEARCH ^\A(?s)(?:.+?^=+){5}.+?\K^=+

Note that the similar regex ^\A(?s)(?:.+?^=+){5}.+?\K^=+\r\n would match every multiple-of-six ====== line ( So the 6th, 12th, 18th line and so on… )

Alan Kilborn

@guy038 said in Add line of text to beginning of multiple files:

I proposed the more rigorous regex (?<!\n|\r|\f)^, because the dot regex does not match the FF control character !

I don’t see the connection here…where is the dot regex being used?

Alan Kilborn

@guy038

In case of your replacement regex ends with … \r\n, it/they must not begin with one/several true empty lines !

I’m not sure I’m following this, either.

How about some real examples that you are usually so good at? :-)

Alan Kilborn

@guy038

Also, an ideas as to why this ^\A seems to work for the intended purpose of matching only at the beginning of a file?

I understand why the earlier (?<!\x0A)^ or (?<!\n|\r|\f)^ work, but ^\A is hard to get one’s head around.

guy038

Hi, @alan-kilborn and All,

Indeed, as the \A is not working properly, a lot of particular cases occur and I confess that detailing them all is a bit tedious work !

Regarding your first question :

I don’t see the connection here…where is the dot regex being used?

The fact of the simple . regex does not match the FF character, of Unicode code-point \x{000C} means that, it is implicitly seen as a non-standard character, i.e. a pseudo EOL character !

So, from the @Terry-r’s solution (?<!\x0A)^, which can be rewritten as (?<!\n)^, I first added the \r case to the negative look-behind, to handle MAC files and then added the Form Feed character case in case some of these chars are present in current file giving the final form (?<!\n|\r|\f)^

For instance, assuming the text, below, where three Form Feed chars ( \x{000c} ) are inserted between the strings 12345 and 67890 of the first line

1234567890
This is a test
to verify
-----><-----
some oddities
on this very
simple text

Paste this text in a new tab
Now, open the Replace dialog
SEARCH (?<!\n|\r)^ ( The \f is not present )
REPLACE ABC
Tick the Wrap around option
Select the Regular expression search mode
Put the caret, on the line -----><----- between symbols > and <
Click on the Replace All button

You get the text :

ABC12345ABCABCABC67890
This is a test
to verify
-----><-----
some oddities
on this very
simple text

As you can see, it correctly adds the string ABC at beginning of file, but also after each FF character :-(

Now, change the search regex as (?<!\n|\r|\f)^
Put the caret, on the line -----><----- between symbols > and <
Click on the Replace All button

This time, we have the expected text, with the string ABC , inserted at beginning if the first line, only ;-))

ABC1234567890
This is a test
to verify
-----><-----
some oddities
on this very
simple text

So, the 3 syntaxes (?<!\n|\r|\f)^, \^A and also \A^ are correct work-arounds to the bugged \A assertion, used alone !

Note, Alan, that the syntax \A^ seems a bit more logic than the opposite form ^\A. However, remark that two consecutive assertions are commutative. For instance, the regexes ^$ and $^ are equivalent as, both, searches for a location which is, either, a beginning and an end of lines and match the zero-length location at the beginning of any empty line

Similarly, two consecutive look-around, which can be considered as user’s assertions, are commutative too ! For instance, the regexes (?-s)(?=ABC)(?=.*XYZ).+ and (?-s)(?=.*XYZ)(?=ABC).+ are equivalent and match any non-null range of chars, beginning with ABC, till the end of current line, if, in addition, the string XYZ occurs from right after ABC till the end of current line !

I hope that these lines answer to your third question :

I understand why the earlier (?<!\x0A)^ or (?<!\n|\r|\f)^ work, but ^\A is hard to get one’s head around

Now, contrary to the 3 correct regex syntaxes, above, the initial regex S/R, against our sample text :

SEARCH \A

REPLACE ABC

gives the complete mess, below, adding the string ABC after any existing character :-((

ABC1ABC2ABC3ABC4ABC5ABCABCABCABC6ABC7ABC8ABC9ABC0ABC
ABCTABChABCiABCsABC ABCiABCsABC ABCaABC ABCtABCeABCsABCtABC
ABCtABCoABC ABCvABCeABCrABCiABCfABCyABC
ABC-ABC-ABC-ABC-ABC-ABC>ABC<ABC-ABC-ABC-ABC-ABC-ABC
ABCsABCoABCmABCeABC ABCoABCdABCdABCiABCtABCiABCeABCsABC
ABCoABCnABC ABCtABChABCiABCsABC ABCvABCeABCrABCyABC
ABCsABCiABCmABCpABClABCeABC ABCtABCeABCxABCtABC
ABC

Now, regarding your second question :

I’m not sure I’m following this, either.

How about some real examples that you are usually so good at? :-)

Well, if we still assume our sample text :

1234567890
This is a test
to verify
-----><-----
some oddities
on this very
simple text

Let’s use, now, the following regex S/R, with Wrap around, Regular expression set and the caret between the >< string :

SEARCH \A^

REPLACE ABC\r\n ( Note that we add some EOL chars, at the end of the replacement regex )

As expected, after clicking on the Replace All button, we obtain the text, with the line ABC inserted before the present first line

ABC
1234567890
This is a test
to verify
-----><-----
some oddities
on this very
simple text

Now, hit Ctrl + Z to undo the changes
Hit 3 times on the Enter key, to create three empty lines, beginning current file
Re-run the same S/R

This time, we get :

ABC

ABC

ABC

ABC
1234567890
This is a test
to verify
-----><-----
some oddities
on this very
simple text

Which is not, obviously, what it is expected ! This is the reason why I said :

In case of your replacement regex ends with the forms \r\n, \n or \r, depending of your file(s) structure, as above, the file(s) must not begin with one/several true empty lines !

Finally :

Hit, again, Ctrl + Z to undo the changes
Use the logical regex S/R, below, with \A, against our sample text, with the first three empty lines :

SEARCH \A

REPLACE ABC\r\n

This time, it’s even worse ! As you can see, below , everything seems to go wrong for all ;-(((

ABC

ABC

ABC

ABC
1ABC
2ABC
3ABC
4ABC
5ABC
ABC
ABC
ABC
6ABC
7ABC
8ABC
9ABC
0ABC

ABC
TABC
hABC
iABC
sABC
 ABC
iABC
sABC
 ABC
aABC
 ABC
tABC
eABC
sABC
tABC

ABC
tABC
oABC
 ABC
vABC
eABC
rABC
iABC
fABC
yABC

ABC
-ABC
-ABC
-ABC
-ABC
-ABC
>ABC
<ABC
-ABC
-ABC
-ABC
-ABC
-ABC

ABC
sABC
oABC
mABC
eABC
 ABC
oABC
dABC
dABC
iABC
tABC
iABC
eABC
sABC

ABC
oABC
nABC
 ABC
tABC
hABC
iABC
sABC
 ABC
vABC
eABC
rABC
yABC

ABC
sABC
iABC
mABC
pABC
lABC
eABC
 ABC
tABC
eABC
xABC
tABC

ABC

Best Regards,

guy038

Alan Kilborn

@guy038 said in Add line of text to beginning of multiple files:

In case of your replacement regex ends with the forms \r\n, \n or \r, depending of your file(s) structure, as above, it/they must not begin with one/several true empty lines !

So I found myself with just such a case. Hmmm, what to do…?

Here’s what I wanted:

Add new first line\r\n to a bunch of files as, obviously, the new first line data.

Here’s what I did:

I did a two-step replacement; first I did:

Find what box: ^\A
Replace with box: new first lineFIRSTLINEPSEUDOENDING

Secondly,

Find what box: FIRSTLINEPSEUDOENDING
Replace with box: \r\n

Of course FIRSTLINEPSEUDOENDING could be any unique string that doesn’t already appear in the files I’m applying this to.

sean o sullivan

Hi All,

i’m trying to add the below 6 lines into the top of 300 batch scripts, but none of the above worked

@echo off
call D:\Tidal\bin\SetupEnv.bat
if errorlevel 1 (
echo Could not retrieve environment variables.
exit /b 1
)

Alan Kilborn

@sean-o-sullivan said in Add line of text to beginning of multiple files:

none of the above worked

What exactly did you try?

Additional question: Do all/some/none of the files you want to affect have empty line(s) at the beginning?

PeterJones

@sean-o-sullivan,

none of the above worked

Notepad++ search/replace boxes use a single line for inputting the expression, so if you want to insert multiple lines, you have to encode the line endings in a way the search engine understands: since you are likely using Windows CRLF line endings for a batch file, every newline in the text you want to insert needs to be converted to \r\n; further, any special character (like the \) needs to be escaped. So, to give an example, the first two lines of your insertion would be something like @echo off\r\ncall D:\\Tidal\\bin\\SetupEnv.bat\r\n, and you’d have to prepare the rest of the lines similarly.

But really, when you’re getting to multi-line-text inserts and hundreds of files, you’re starting to move beyond what is the primary focus of text editor software: it focuses more on one or a few files, rather than bulk editing (though Find in Files is a nice move toward bulk-file-editing).

Operating systems like Windows and Linux are primarily GUI nowadays, but they still ship with one or more command line systems; there’s a reason for that: you can do powerful things at the command line, often on lots of files in a short amount of time, with very little code.

Backup any critical batch files: your files are always your responsibility
Create a file called prefix.in, which has your 6 lines that you want to insert (make sure to have a NEWLINE at the end of the last line), and save.
Put prefix.in in the same directory as the three-hundred *.bat files
open cmd.exe and change to that directory
for %F in (*.bat) do @( copy prefix.in + %F %F.out & move /Y %F.out %F )

This will look up the name (%F) of every batch file in the folder, it will concatenate the contents of prefix.in and %F into a new temporary file %F.out, then it will rename %F.out back to %F. (I would have used REN or RENAME instead of MOVE, but REN doesn’t have the /Y to force allowing overwriting of the destination.)

With one line of command-line syntax, I just prefixed all 300 batch files with the 6 lines you wanted.

(Whenever I need to look up cmd.exe syntax, I use https://ss64.com/nt/syntax.html for topic-level, and the “CMD” link https://ss64.com/nt/ for a per-commmand reference. I know, using cmd.exe is old fashioned, but I’ve never gotten the hang of powershell.)

Notepad++ is great, but there are other tools available to you.

Alan Kilborn

@PeterJones does the legwork I was trying to get the OP to do. :-(

I think it’s an OK task for Notepad++. A lot of people are scared to dip down into the world of batch/CMD/Powershell…

The OP even did the research to find this thread. However, the OP should have done more than “Wah! It doesn’t work”. Not a direct quotation. :-)

Notepad++ …/replace boxes use a single line for inputting the expression, so if you want to insert multiple lines, you have to encode the line endings in a way the search engine understands: since you are likely using Windows CRLF line endings for a batch file, every newline in the text you want to insert needs to be converted to \r\n; further, any special character (like the ) needs to be escaped.

People are scared also to do Pythonscript, so I hesitate to bring this up, but a script I use can do the conversion of the data per the above. It takes the currently selected text, formats it correctly, and puts the result in the clipboard, ready for pasting into the Replace with box:

# -*- coding: utf-8 -*-

from Npp import editor, notepad

class T11987(object):
    def __init__(self):
        if editor.getSelections() > 1: return  # acting on multiple selections or a column block probably doesn't make sense
        s = editor.getSelText()
        if len(s) == 0: return
        new_s = s.replace('\\', '\\\\').replace('\r', '\\r').replace('\n', '\\n').replace('(', '\\(').replace(')', '\\)')
        if len(new_s) > 2046:
            notepad.messageBox('After conversion, the selected text is too long for the ***Replace with*** box (limit: ~2046 chars)', '')
        else:
            editor.copyText(new_s)
            notepad.messageBox('The selected text has been converted and is now in the clipboard.', '')

if __name__ == '__main__': T11987()

Running the script on the OP’s replace data yields the full string that Peter started:

@echo off\r\ncall D:\\Tidal\\bin\\SetupEnv.bat\r\nif errorlevel 1 $\r\necho Could not retrieve environment variables.\r\nexit /b 1\r\n$\r\n

There might be more characters that need escaping in a regex replace string, but in practice I haven’t encountered them yet.

sean o sullivan

@PeterJones You sir are a legend, that worked perfectly!!! Thank you so much for taking the time to help me out

Alan Kilborn

A lot of people are scared to dip down into the world of batch/CMD/Powershell…

@PeterJones You sir are a legend, that worked perfectly!!!

…and some people are not afraid. :-)

One point of clarification on my previous reply:

@PeterJones had previously stated:

Notepad++ search/replace boxes use a single line for inputting the expression

but my reply concentrated only on the replace box. The reason for that is that if you want to get multiline text that you already have in the Notepad++ window into the “search box”, all you have to do is select it and press ctrl+f (if ctrl+f is so configured – to copy into the box automatically). No conversion of the text (or escaping) is necessary.

UNLESS, you are going to be doing a multiline replacement – in which case you need to be in Regular expression search mode – and your search data contains any of the special regex metacharacters. In this case the best thing to do might be to add a \Q at the very start of the “search box”.

Yikes, this might be really “obscure”!

PeterJones

@sean-o-sullivan ,

Glad you got something that worked. I hope you took the opportunity to learn from what @Alan-Kilborn shared as well, because it really is helpful search/replace advice, too.

-----

@Alan-Kilborn said in Add line of text to beginning of multiple files:

I think it’s an OK task for Notepad++.

Well, the 6 lines aren’t bad to encode into the replace box. But when the next poster wants to insert 1023 lines at the beginning of every config file? As I said, “you’re starting to move beyond” (emphasis added): inserting 6 lines at the beginning of 300 files has started down the slippery slope out of the realm of pure text-editor, but not so far that it’s not doable.

Pythonscript, … It takes the currently selected text, formats it correctly, and puts the result in the clipboard, ready for pasting into the Replace with box

I didn’t read correctly / realize what you were doing there until the second read: that’s just cool.

A side question, unrelated to OP: why create an object with the __init__, rather than just have a T11987() function instead?

my reply concentrated only on the replace box. … No conversion … necessary

True enough.

Alan Kilborn

@PeterJones said in Add line of text to beginning of multiple files:

inserting 6 lines at the beginning of 300 files has started down the slippery slope out of the realm of pure text-editor

Agreed. Maybe that’s why I said it’s an “OK” task for Notepad++ and not “a great” task. :-)

why create an object with the init, rather than just have a T11987() function instead?

It’s just my new framework for scripts (I copy from some boilerplate when I make a new script). Obviously in this case it is fairly trivial, but in scripts that use callbacks a object-based approach has bigger benefits. I picked up on this by observing how @Ekopalypse does some of the more complicated scripts he’s presented. BTW, I picked up on using the topic/posting id in the name of a script/function/class from YOU. :-)