How to extract part of a document and save it to a new one (command line Mac)

Question

Our buildserver needs to release apps to App Center. App Center allows me to provide a release note file.

The problem is, our ReleaseNotes.md contains ALL the versions and become to big to for App Center to accept it.

The notes are formatted like this:

# Project - Release Notes - Android

## 1.20.2 - 2020-09-11
* [UID-4782] - Connectivity issues fixed

## 1.20.1 - 2020-09-08
 * [UID-4639] - Update Color
 * [UID-4760] - Changed some stuff

How can I grab just the first entry and save it to a file?:

## 1.20.2 - 2020-09-11
 * [UID-4782] - Connectivity issues fixed

and save it to a file?

If looked into tools like awk, grep, sed and pcergrep but I'm not familiar with these in anyway and I have no idea which one would be the right tool for the job.

anubhava · Answer 1 · 2020-10-20T14:45:07.053

2

This awk should get the job done:

awk -v RS= '/^##/{print; exit}' file

## 1.20.2 - 2020-09-11
* [UID-4782] - Connectivity issues fixed

edited Oct 20 '20 at 14:45

answered Oct 19 '20 at 15:36

anubhava

761,203
64
569
643

For me it resulted in just: ```##``` – SEG.Veenstra Oct 20 '20 at 14:29
1

Thank you, this also works! Now I have multiple solutions that work! – SEG.Veenstra Oct 20 '20 at 14:47

score 1 · Answer 2 · answered Oct 19 '20 at 15:28

1

Using awk in "paragraph mode" by setting RS to the empty string and then saving it to another file:

awk -v RS= '/issues fixed/' file > file2
## 1.20.2 - 2020-09-11
* [UID-4782] - Connectivity issues fixed

answered Oct 19 '20 at 15:28

Carlos Pascual

1,106
1
5
8

1

If I understand this correctly it is depending on the `'issues fixed'` right? But I also need it to work if a new entry is being added which might not contain that. – SEG.Veenstra Oct 20 '20 at 14:31
Note that you get the whole paragraph. You can experiment putting something different present in the paragraph, for example /2020-09-11/ You'll get the same, i.e., the whole paragraph. – Carlos Pascual Oct 20 '20 at 17:46
Ah yes, I see, but now it's giving me all the paragraphs. I only want the first one. (without knowing what is in it) – SEG.Veenstra Oct 21 '20 at 06:36

score 1 · Accepted Answer · edited Oct 20 '20 at 14:42

1

Using sed:

sed -n '/^##/,/^$/{p;/^$/q;}' input_file > output_file

-n: Do not print lines by default.
/^##/: Find a line starting with ##.
/^$/: Find an empty line.
/^##/, /^$/: Together this means for each line between a line starting with ## and an empty line.
p: Print the line.
/^$/q: If the line is empty quit. We do this because otherwise sed will process the next block of changes.

edited Oct 20 '20 at 14:42

SEG.Veenstra

718
1
7
18

answered Oct 20 '20 at 13:57

perreal

94,503
21
155
181

It is giving me the following output: `sed: 1: "/^##/,/^$/{p;/^$/q}": extra characters at the end of q command` – SEG.Veenstra Oct 20 '20 at 14:12
Hmm, strange. see: https://sed.js.org/?gist=532d72251c4cb7869f19e34f4de301c4#. What is your OS? sed version? – perreal Oct 20 '20 at 14:29
MacOS Catalina, I can't find sed's version number – SEG.Veenstra Oct 20 '20 at 14:35
1

Nice, solved it, I had to put a `;` behind the `q` it works! – SEG.Veenstra Oct 20 '20 at 14:38
Accepted as the answer because it works, and has the best explanation. – SEG.Veenstra Oct 20 '20 at 14:44

SEG.Veenstra · Answer 4 · 2020-10-20T15:09:04.263

1

So I've did some bash script learning and created the following script:

#!/bin/bash

foundFirstEntry=false

"ReleaseNotes.md" | while read p; do
    if [[ $p = \#\#* ]]
    then
        foundFirstEntry=true
    fi

    if [[ $foundFirstEntry = true && $p = "" ]]
    then
        break
    fi

    if $foundFirstEntry 
    then
        echo "$p"
    fi
done > "ShortReleaseNotes.md" < "ReleaseNotes.md"

edited Oct 20 '20 at 15:09

answered Oct 20 '20 at 14:22

SEG.Veenstra

718
1
7
18

1

That's great. You can remove the cat and change the last line to `done > "ShortReleaseNotes.md" < "ReleaseNotes.md"`. – perreal Oct 20 '20 at 14:41
Is that better? If so, can you explain why? – SEG.Veenstra Oct 20 '20 at 14:45
1

It is better because `cat` needs to be loaded and executed but it is not necessary. see: https://stackoverflow.com/questions/11710552/useless-use-of-cat. – perreal Oct 20 '20 at 14:51

How to extract part of a document and save it to a new one (command line Mac)

4 Answers4