How to split git log output into an array of commits in bash?

Question

I'm not good at bash, and what I was trying to do is to get information from git log --name-status branch_one...branch_two and after that split output by commits, for example:

git_command_output="
commit one
Author: Johny
Date: 01/01/1999

feat: add something

M src/some_file.txt

commit two
Author: Alex
Date: 02/01/1999

fix: bug

M src/some_file.txt"

From the above I would want to create and array of strings, where each string will be information about one commit.

For instance array would look something like that:

echo "${array[0]}" 

# Author: Johny
# Date: 01/01/1999

# feat: add something

# M src/my_file.txt

echo "${array[1]}" 

# Author: Alex
# Date: 02/01/1999

# fix: bug

# M src/foreign_file.txt

What I tried (didn't work):

array[0]=$(echo "$git_command_output" | awk '{split($0, array, "commit"); print array[0]}')

IFS='commit' read -r -a array <<< "$git_command_output"

Is there a concise way to do it, preferably using awk?

anubhava · Accepted Answer · 2022-10-04 11:19:53Z

4

You may use this readarray with awk:

IFS= readarray -d '' -t arr < <(awk -v RS='\nM [^\n]+\n' '{ORS = RT "\0"} 1' file)

# check array content
declare -p arr

declare -a arr=([0]=$'commit one\nAuthor: Johny\nDate: 01/01/1999\n\nfeat: add something\n\nM src/some_file.txt\n' [1]=$'\ncommit two\nAuthor: Alex\nDate: 02/01/1999\n\nfix: bug\n\nM src/some_file.txt\n')

Here:

We add \0 after each record using awk
Then use -d '' to make sure to read \0 delimited text in array arr

Another simpler any-version awk would be:

IFS= readarray -d '' -t arr < <(awk '{printf "%s", $0 (/^M / ? "\0" : ORS)}' file)

edited Oct 4, 2022 at 11:19

answered Oct 3, 2022 at 16:18

anubhava

790k67 gold badges603 silver badges671 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Slava.In Over a year ago

I'm experimenting now, for some reason "${arr[0]}" shows the whole contents, while "${arr[1]}" shows none

anubhava Over a year ago

Please check your awk and bash versions. It requires gnu-awk and bash 4+

Slava.In Over a year ago

Is there a way to do the same thing, but read from a string variable instead of a file? -

anubhava Over a year ago

Yes sure, just replace awk -v RS='\nM [^\n]+\n' '{ORS = RT "\0"} 1' file with awk -v RS='\nM [^\n]+\n' '{ORS = RT "\0"} 1' <<< "$string"

1ace · Accepted Answer · 2022-10-03 16:04:12Z

2

You can filter the output of many git commands (such as log) to those that affect specific files, simply by adding the paths of these files at the end of the command. Try:

git log --name-status branch_one...branch_two -- src/my_file.txt

answered Oct 3, 2022 at 16:04

1ace

5,2984 gold badges26 silver badges30 bronze badges

Collectives™ on Stack Overflow

How to split git log output into an array of commits in bash?

2 Answers 2

4 Comments

Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Related