Ayende @ Rahien

It's a girl

Extracting a list of committers from Git

There doesn’t seem to be a nice way to getting stats like “who are the committers in this project” from Git. There is probably some fancy script that does sed/awk magic to do so, but I went with a simpler alternative:

git log --raw > log.txt

var logFile = @"C:\work\ravendb\log.txt";

var committers = from line in File.ReadAllLines(logFile)
                 where line.StartsWith("Author: ")
                 let author = line.Substring("Author: ".Length)
                 group author by author
                 into g
                 let result = new {Committer = g.Key, CommitsCount = g.Count()}
                 orderby result.CommitsCount descending
                 select result;

foreach (var committer in committers)
{
    Console.WriteLine(committer);
}

Running this on Raven’s code produces:

{ Committer = Ayende Rahien <ayende>, CommitsCount = 555 }
{ Committer = Ayende Rahien <Ayende>, CommitsCount = 477 }
{ Committer = unknown <Ayende, CommitsCount = 72 }
{ Committer = Paul B <github, CommitsCount = 24 }
{ Committer = Andrew Stewart <Andrew.Stewart, CommitsCount = 24 }
{ Committer = Benny Thomas <jan.thomas, CommitsCount = 19 }
{ Committer = Luke Hertert <lukehertert, CommitsCount = 15 }
{ Committer = Bobby Johnson <bobby.johnson, CommitsCount = 13 }
{ Committer = Rob Ashton <robashton, CommitsCount = 11 }
{ Committer = unknown <LukeHertert, CommitsCount = 7 }
{ Committer = Andrew Theken <theken.1, CommitsCount = 6 }
{ Committer = AndyStewart <Andy.Stewart, CommitsCount = 3 }
{ Committer = Steve Strong <steve, CommitsCount = 3 }
{ Committer = unknown <Aaron Weiker, CommitsCount = 1 }
{ Committer = Emil Cardell <emil.cardell, CommitsCount = 1 }
{ Committer = bjarte.skogoy <bjarte.skogoy, CommitsCount = 1 }
{ Committer = unknown <henke, CommitsCount = 1 }

Comments

Martin Larsen
06/01/2010 10:29 AM by
Martin Larsen

git log --raw | grep "^Author: " | sort | uniq -c

Masklinn
06/01/2010 10:44 AM by
Masklinn

There doesn’t seem to be a nice way to getting stats like “who are the committers in this project” from Git.

The gitstat tool from the internets looks nice, considering git apparently doesn't have the equivalent of hg churn.

Frank Quednau
06/01/2010 12:02 PM by
Frank Quednau

For completeness a powershell version that uses what is contained in the < and >:

git log --raw | Select-String -Pattern "^Author.<(.)>" | % { $_.Matches[

0].Groups[1].Value } | Group-Object | select count, name

Chris
06/01/2010 06:13 PM by
Chris

+1 Larsen's solution

Ayende Rahien
06/01/2010 06:59 PM by
Ayende Rahien

Just one problem with all of the things that were suggested.

They requires me to THINK about those.

My solution was coded under heavy time pressure and I just needed the answers.

Going with the route of least resistance was the easiest and faster way to get things done.

Johannes Rudolph
06/01/2010 10:20 PM by
Johannes Rudolph

That's why I couldn't live without linqPad anymore!

Martin Larsen
06/02/2010 05:47 AM by
Martin Larsen

Ah... Cosmin 0wns us all. :-)

Marcin Floryan
06/14/2010 02:13 PM by
Marcin Floryan

Nice one, but how about streaming from a file? You'd be better off not having to read the whole file into memory which is precisely what LINQ allows you to do.

public IEnumerator

<string GetEnumerator()

    {

        using (TextReader reader = dataSource())

        {

            string line;

            while ((line = reader.ReadLine()) != null)

            {

                yield return line;

            }

        }

    }
Comments have been closed on this topic.