Insert string/words above to lines contain numbers

Isaac Goh

I need to convert 5000+ lines as shown below. Basically, need to insert string/word from above to lines containing number below till before the next string/word. Thanks in advanced.

Convert from this:
Orange
50
20
apple
30
20
10
pear
50
20
21
10

To this:
Orange 50
Orange 20
apple 30
apple 20
apple 10
pear 50
pear 20
pear 21
pear 10

guy038

Hello, @isaac-goh and All,

No problem with regexes ! we’ll need 2 consecutive regex S/R. As usual :

Move to the very beginning of your file ( Ctrl + Home )
Select the Regular expression search mode
Untick all box options
Click once only on the Replace All button

So, from your INPUT text :

Orange
50
20
apple
30
20
10
pear
50
20
21
10

This first regex S/R

SEARCH ^([\u\l]+\R)((?:\d+\R)+)

REPLACE \2\1

will reverse the position of all the articles and their respective quantities :

50
20
Orange
30
20
10
apple
50
20
21
10
pear

Then, this second regex will give the expected OUTPUT text by pre-fixing any quantity with its article name :

SEARCH ^(\d+\R)(?=(?s:.*?)([\u\l]+))|^[\u\l]+\R

REPLACE ?1\2\x20\1

Orange 50
Orange 20
apple 30
apple 20
apple 10
pear 50
pear 20
pear 21
pear 10

In addition, these replacements are quite safe. Indeed, performing them, against the OUPUT text only, will not modify anything else ;-))

Best Regards,

guy038

Isaac Goh

Hi guy038,
Thanks for your prompt post. But I don’t seem to get your post to work

For the first regexes
SEARCH ^([\u\l]+\R)((?:\d+\R)+)
REPLACE \2\1

The outout is a bit out:
50
20
Orange
30
20
10
apple
50
20
21
pear
10

Thus the second regexes
SEARCH ^(\d+\R)(?=(?s:.*?)([\u\l]+))|^[\u\l]+\R
REPLACE ?1\2\x20\1

Orange 50
Orange 20
apple 30
apple 20
apple 10
pear 50
pear 20
pear 21
10

PeterJones

@Isaac-Goh said in Insert string/words above to lines contain numbers:

But I don’t seem to get your post to work

That’s because your data doesn’t have a final newline after the 10. Guy’s regex assumes that every line that is a number ends with a newline.

Your choices are to add a newline to the end before running the two regex, or replace the first regex with ^([\u\l]+\R)((?:\d+(\R|\Z))+)

----

Useful References

PeterJones

@Isaac-Goh ,

Alternatively,

First:
FIND = ^([\u\l]+)\R
REPLACE = $1\x20
Replace All
Second:
FIND = ^([\u\l]+)\x20\d+\R\K(\d+)$
REPLACE = $1\x20$2
Replace All one or more times until the message says none were replaced

guy038

Hello, @isaac-goh, @peterjones and All,

Peter, thanks for your clever alternative solution ! I didn’t think of that,. Of course, it needs as many clicks on the Replace All button than the greatest list of numbers minus 1 !

Now, your solution to handle a last line without any [CR/]LF , that you gave to @isaac-goh :

^([\u\l]+\R)((?:\d+(\R|\Z))+)

does not suit too :-(

Indeed, you get this OUTPUT text :

50
20
Orange
30
20
10
apple
50
20
21
10pear

So, in summary, in order to solve the @isaac-goh goal :

My first regex S/R should have been :
- SEARCH ^([\u\l]+\R)((?:\d+(?:\R|(\Z)))+)
- REPLACE $2(?3\r\n)$1 or $2(?3\n)$1 if you use Unix files
My second regex S/R is unchanged :
- SEARCH ^(\d+\R)(?=(?s:.*?)([\u\l]+))|^[\u\l]+\R
- REPLACE ?1$2\x20$1

Best Regards,

guy038

PeterJones

@guy038 said in Insert string/words above to lines contain numbers:

Of course, it needs as many clicks on the Replace All button than the greatest list of numbers minus 1

That doesn’t bother me. I’d much rather spend a few extra seconds clicking Replace All a few more times than spend XYZ more minutes making the “perfect” regex that does it all with one click. I’ll put just enough effort into a regex for Notepad++ so that it gets the job done and that I can still understand it (and can hope that I will understand it if I read it again six months down the line when I want to try something similar again). Usually, the extra XYZ minutes to find some “clever trick” just ends up making the regex so complicated that I’d never remember the trick in the future, and might not even be able to see what the regex does by looking at it. And if I cannot see what it does, I’ll never try using it again in the future.

Now, your solution to handle a last line without any [CR/]LF , that you gave to @isaac-goh :
^([\u\l]+\R)((?:\d+(\R|\Z))+)
does not suit too :-(

I could’ve sworn I tried it yesterday when I posted. But when I tried it this morning, I got the same results as you. (My guess is that the final EOL had been added into my document without me noticing when I tried my regex yesterday…) So thanks for that correction.

Isaac Goh

@guy038

Your details and prompt effort are very much appreciated. Thanks a million

I realise that if certain lines contain as below. The regex doesn’t seem to work.
My apology for not listing it in my initial post. Anyway, thank you guys for the great effort.

Orange123
50:00
20:00
apple234
30:00
20:00
10:00
pear456
50:00
20:00
21:00
10:00

guy038

Hi, @isaac-goh ,

Well, as usual , from this INPUT text :

Orange123
50:00
20:00
apple234
30:00
20:00
10:00
pear456
50:00
20:00
21:00
10:00

Which OUTPUT do you expect ?

BR

guy038

Isaac Goh

@guy038
Output will be as shown below. Many thanks

Orange123 50:00
Orange123 20:00
apple234 30:00
apple234 20:00
apple234 10:00
pear456 50:00
pear456 20:00
pear456 21:00
pear456 10:00

guy038

Hello, @isaac-goh, @peterjones and All,

Ah OK ! So, now, we have to find out something less restrictive, because :

The articles may contain digits or an underscore in their name
The quantities may contain the symbol :

In order that the regexes clearly recognize these two entities, I assume that the articles will always begin with an upper or lower letter !

So, from this INPUT text :

Orange123
50:00
20:00
apple234
30:00
20:00
10:00
pear456
50:00
20:00
21:00
10:00

The first regex S/R to use becomes :

SEARCH ^([\u\l]\w*\R)((?:[\d:]+(?:\R|(\Z)))+)
REPLACE $2(?3\r\n)$1 or $2(?3\n)$1 if you use Unix files

and gives :

50:00
20:00
Orange123
30:00
20:00
10:00
apple234
50:00
20:00
21:00
10:00
pear456

Then, the second regex S/R to use becomes :

SEARCH ^([\d:]+\R)(?=(?s:.*?)([\u\l]\w*)$)|^[\u\l]\w*\R
REPLACE ?1$2\x20$1

and gives the expected OUTPUT text :

Orange123 50:00
Orange123 20:00
apple234 30:00
apple234 20:00
apple234 10:00
pear456 50:00
pear456 20:00
pear456 21:00
pear456 10:00

Jut tell me if additional symbols must appear, apart from the : character, in the quantities !

Best Regards,

guy038

Isaac Goh

@guy038
Yes additional symbols like underscore _ and additional colon : as shown below. Many thanks.

Convert from this:
Orange123_432
50:00:00
20:00:00
apple234_678
30:00:00
20:00:00
10:00:00
pear456_321
50:00:00
20:00:00
21:00:00
10:00:00

To this:
Orange123_432 50:00:00
Orange123_432 20:00:00
apple234_678 30:00:00
apple234_678 20:00:00
apple234_678 10:00:00
pear456_321 50:00:00
pear456_321 20:00:00
pear456_321 21:00:00
pear456_321 10:00:00

Best Regards

Isaac Goh

My apology as there are more than 3000+ lines. I realised additional symbols like underscore _ additional colon : and alphanumeric as shown below. Many thanks.

Convert from this:
Orange123_432
50:00:00:1e
20:00:00:1e
apple234_678
30:00:00:1a
20:00:00:1a
10:00:00:1a
pear456_321
50:00:00:1b
20:00:00:1b
21:00:00:1b
10:00:00:1b

To this:
Orange123_432 50:00:00:1e
Orange123_432 20:00:00:1e
apple234_678 30:00:00:1a
apple234_678 20:00:00:1a
apple234_678 10:00:00:1a
pear456_321 50:00:00:1b
pear456_321 20:00:00:1b
pear456_321 21:00:00:1b
pear456_321 10:00:00:1b

Best Regards

PeterJones

@Isaac-Goh ,

So what is the rule you use to decide which is the “header” line and which is the “other” lines? Is the rule “if it starts with a letter (uppercase or lowercase), it is a header line; if it starts with a number, it is an ‘other’ line”? Or something else.

You keep on changing what you expect. If you do not give a reasonable definition of what you want, you cannot expect that someone would get it right for you.

----

Useful References

guy038

Hi, @isaac-goh, @peterjones and All,

Well. So :

The articles contain letters, may contain digits or an underscore in their name and begin with a letter
The quantities contains digits, may contain the : symbol and may have a trailing lower-case letter

In this case, use successively these two regex S/R, below :

SEARCH ^([\u\l]\w*\R)((?:[\d:]+\l?(?:\R|(\Z)))+)
REPLACE $2(?3\r\n)$1 or $2(?3\n)$1 if you use Unix files

and :

SEARCH ^([\d:]+\l?\R)(?=(?s:.*?)^([\u\l]\w*)$)|^[\u\l]\w*\R
REPLACE ?1$2\x20$1

Which will change the INPUT test :

Orange123_432
50:00:00:1e
20:00:00:1e
apple234_678
30:00:00:1a
20:00:00:1a
10:00:00:1a
pear456_321
50:00:00:1b
20:00:00:1b
21:00:00:1b
10:00:00:1b

into the expected OUTPUT text :

Orange123_432 50:00:00:1e
Orange123_432 20:00:00:1e
apple234_678 30:00:00:1a
apple234_678 20:00:00:1a
apple234_678 10:00:00:1a
pear456_321 50:00:00:1b
pear456_321 20:00:00:1b
pear456_321 21:00:00:1b
pear456_321 10:00:00:1b

BR

guy038

Isaac Goh

@guy038
Apology for my late reply as I am out of town for almost a week and need to scan thru more than 3000+ lines. This is what I have gathered as the possible longest lines.

Isaac Goh

@guy038

Convert from this:
hq12orange1_HBA0_Max_123_PG1_FA1D30
21:00:00:0b:1f:c1:f6:ee
50:00:09:1a:50:06:1b:1f
10:00:00:10:ca:b1:60:94
H600_654342_3B_hq11apple6_HBA4
50:06:0b:50:12:45:fa:20
21:00:12:40:0d:c2:4b:75
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1
50:00:15:42:c1:32:21:1a
50:00:19:45:d1:06:14:04

To this:
hq12orange1_HBA0_Max_123_PG1_FA1D30 21:00:00:0b:1f:c1:f6:ee
hq12orange1_HBA0_Max_123_PG1_FA1D30 50:00:09:1a:50:06:1b:1f
hq12orange1_HBA0_Max_123_PG1_FA1D30 10:00:00:10:ca:b1:60:94
H600_654342_3B_hq11apple6_HBA4 50:06:0b:50:12:45:fa:20
H600_654342_3B_hq11apple6_HBA4 21:00:12:40:0d:c2:4b:75
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1 50:00:15:42:c1:32:21:1a
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1 50:00:19:45:d1:06:14:04

Best Regards

Alan Kilborn

@Isaac-Goh

So this isn’t a “data conversion service”.
The way it works is you get a couple of “freebies” and then that stops and you are expected to learn from what you have been given.
Of course you may ask questions about any small points, but in general you are expected to do your own work.
Also you may show what you’ve tried and explain how it wasn’t successful and ask for hints, but you may no longer ask for full solutions.

guy038

Hi, @isaac-goh, @peterjones, @alan-kilborn and All,

Of course, @alan-kilborn’s comments are pertinent and you should investigate a bit in regexes : refer to the USEFUL REFERENCES part at the end of the first post of @peterjones !

Now, out of curiosity, I wanted to know, why you asked me, again. Seemingly, my provided regexes did not match all your cases !

Oh…, everything is clear, now ! Really, @isaac-goh, you should have told me, from the very beginning that :

The header lines are a range of word characters \w, so digits, letters, whatever its case or underscore )
Each subsequent line is a range of hexadcimal digits [[:xdigit:]], separated with a colon char

This completly change the regexes, of course !

So from this INPUT text :

hq12orange1_HBA0_Max_123_PG1_FA1D30
21:00:00:0b:1f:c1:f6:ee
50:00:09:1a:50:06:1b:1f
10:00:00:10:ca:b1:60:94
H600_654342_3B_hq11apple6_HBA4
50:06:0b:50:12:45:fa:20
21:00:12:40:0d:c2:4b:75
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1
50:00:15:42:c1:32:21:1a
50:00:19:45:d1:06:14:04

with this first regex S/R :

SEARCH ^(\w+\R)((?:[[:xdigit:]:]+(?:\R|(\Z)))+)
REPLACE $2(?3\r\n)$1 or $2(?3\n)$1 if you use Unix files

It’ll gives you this text :

21:00:00:0b:1f:c1:f6:ee
50:00:09:1a:50:06:1b:1f
10:00:00:10:ca:b1:60:94
hq12orange1_HBA0_Max_123_PG1_FA1D30
50:06:0b:50:12:45:fa:20
21:00:12:40:0d:c2:4b:75
H600_654342_3B_hq11apple6_HBA4
50:00:15:42:c1:32:21:1a
50:00:19:45:d1:06:14:04
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1

Then, with the second regex S/R, below :

SEARCH ^([[:xdigit:]:]+\R)(?=(?s:.*?)^(\w+)$)|^\w+\R
REPLACE ?1$2\x20$1

You’ll get the expected OUTPUT text :

hq12orange1_HBA0_Max_123_PG1_FA1D30 21:00:00:0b:1f:c1:f6:ee
hq12orange1_HBA0_Max_123_PG1_FA1D30 50:00:09:1a:50:06:1b:1f
hq12orange1_HBA0_Max_123_PG1_FA1D30 10:00:00:10:ca:b1:60:94
H600_654342_3B_hq11apple6_HBA4 50:06:0b:50:12:45:fa:20
H600_654342_3B_hq11apple6_HBA4 21:00:12:40:0d:c2:4b:75
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1 50:00:15:42:c1:32:21:1a
VICTOR_200_BE_E3A1BC00_MAX_123_FA1D1 50:00:19:45:d1:06:14:04

Best Regards,

guy038

Isaac Goh

@guy038 said in Insert string/words above to lines contain numbers:

?1$2\x20$1

Your superb effort is much appreciated. Many thanks.