How to perform better document version control on Excel files and SQL schema files

Question

I am in charge of several Excel files and SQL schema files. How should I perform better document version control on these files?

I need to know the part modified (different part) in these files and keep all the versions for reference. Currently I am appending the time stamp on the file name, but I found it seemed to be inefficient.

Is there a way or good practice to do better document version control?

By the way, editors send me the files via email.

Answer 1

The answer I have written here can be applied in this case. A tool called xls2txt can provide human-readable output from .xls files. So in short, you should put this to your .gitattributes file:

*.xls diff=xls

And in the .git/config:

[diff "xls"]
    binary = true
    textconv = /path/to/xls2txt

Of course, I'm sure you can find similar tools for other file types as well, making git diff a very useful tool for office documents. This is what I currently have in my global .gitconfig:

[diff "xls"]
    binary = true
    textconv = /usr/bin/py_xls2txt
[diff "pdf"]
    binary = true
    textconv = /usr/bin/pdf2txt
[diff "doc"]
    binary = true
    textconv = /usr/bin/catdoc
[diff "docx"]
    binary = true
    textconv = /usr/bin/docx2txt

The Pro Git book has a good chapter on the subject: 8.2 Customizing Git - Git Attributes

Answer 2

Since you've tagged your question with git I assume you are asking about Git usage for this.

Well, SQL dumps are normal text files so it makes perfect sense to track them with Git. Just create a repository and store them in it. When you get a new version of a file, simply overwrite it and commit, Git will figure out everything for you, and you'll be able to see modification dates, checkout specific versions of this file and compare different versions.

The same is true for .xlsx if you decompress them. .xlsx files are zipped up directories of XML files (See How to properly assemble a valid xlsx file from its internal sub-components?). Git will view them as binary unless decompressed. It is possible to unzip the .xlsx and track the changes to the individual XML files inside of the archive.

You could also do this with .xls files, but the problem here is that .xls format is binary, so you can't get meaningful diffs from it. But you'll still be able to see modification history and checkout specific versions.

Answer 3

I've been struggling with this exact problem for the last few days and have written a small .NET utility to extract and normalise Excel files in such a way that they're much easier to store in source control. I've published the executable here:

https://bitbucket.org/htilabs/ooxmlunpack/downloads/OoXmlUnpack.exe

..and the source here:

https://bitbucket.org/htilabs/ooxmlunpack

If there's any interest I'm happy to make this more configurable, but at the moment, you should put the executable in a folder (e.g. the root of your source repository) and when you run it, it will:

• Scan the folder and its subfolders for any .xlsx and .xlsm files
• Take a copy of the file as *.orig.
• Unzip each file and re-zip it with no compression.
• Pretty-print any files in the archive which are valid XML.
• Delete the calcchain.xml file from the archive (since it changes a lot and doesn't affect the content of the file).
• Inline any unformatted text values (otherwise these are kept in a lookup table which causes big changes in the internal XML if even a single cell is modified).
• Delete the values from any cells which contain formulas (since they can just be calculated when the sheet is next opened).
• Create a subfolder *.extracted, containing the extracted zip archive contents.

Clearly not all of these things are necessary, but the end result is a spreadsheet file that will still open in Excel, but which is much more amenable to diffing and incremental compression. Also, storing the extracted files as well makes it much more obvious in the version history what changes have been applied in each version.

If there's any appetite out there, I'm happy to make the tool more configurable since I guess not everyone will want the contents extracted, or possibly the values removed from formula cells, but these are both very useful to me at the moment.

In tests, a 2 MB spreadsheet 'unpacks' to 21 MB, but then I was able to store five versions of it with small changes between each, in a 1.9 MB Mercurial data file, and visualise the differences between versions effectively using Beyond Compare in text mode.

NB: although I'm using Mercurial, I read this question while researching my solution and there's nothing Mercurial-specific about the solution, should work fine for Git or any other VCS.

Answer 4

Tante recommended a very simple approach in Managing ZIP-based file formats in Git:

Open your ~/.gitconfig file (create if not existing already) and add the following stanza:

[diff "zip"]
textconv = unzip -c -a

Answer 5

Use the open document extension .fods. It's a plain, uncompressed XML markup format that both Excel and LibreOffice can open, and the diffs will look good.

Answer 6

We've built an open-source Git command line extension for Excel workbooks: https://www.xltrail.com/git-xltrail.

In a nutshell, the main feature is that it makes git diff work on any workbook file formats so that it shows the diff on the workbook's VBA content (at some point, we'll make this work for the worksheets content, too).

It's still early days but it might help.

Answer 7

This Excel utility works very well for me:

Version Control for Excel

It is a quite straightforward versioning tool for workbooks and VBA macros. Once you commit a version, it is saved to a Git repository on your PC. I never tried it re. SQL schema files, but I'm sure there's a way around.

Answer 8

My approach with Excel files is similar to Jon's, but instead of working with the raw Excel text data I export to more friendly formats.

Here is the tool that I use: https://github.com/stenci/ExcelToGit/tree/master

All you need is to download the .xlsm file (click the View Raw link on this page.) Don't forget to check the Excel setting as described in the readme. You can also add the code to export SQL data to text files.

The workbook is both a converter from binary Excel to text files and a launcher of the windows Git tools, and it can be used also with non Excel related projects.

My working version is configured with dozens of Excel workbooks. I use the file also to open Git-gui for non Excel projects, just adding the git folder by hand.

Answer 9

As mentioned in another answer's comment, .xlsx files are just XML.

To get to the XML directory (which is git-able), you have to "unzip" the .xlsx file to a directory. A quick way see this on Windows is to rename the file <filename>.xlsx to <filename>.zip, and you'll see the inner contents. I'd store this along with the binary so that when you checkout, you do not have to do other steps in order to open the document in Excel.

Answer 10

If you are not willing to download third-party converters into your project, another option is to write your own script to convert the Excel file for you. I added my own converter using Python. The python script:

from pandas import DataFrame, read_excel
from argparse import ArgumentParser
from os import remove


def change_file_format_to_csv(filename):
    filename = filename.split(".")
    filename[-1] = 'csv'


if __name__ == "__main__":
    # Parse arguments
    parser = ArgumentParser()
    parser.add_argument("-i", "--input", default="", required=False,
                        help="Input file to be converted")
    args = parser.parse_args()

    # Load input
    print(args.input)
    content = read_excel(args.input)

    # Change filename to csv
    filename_output = args.input.split(".")
    filename_output[-1] = "csv"
    filename_output = '.'.join(filename_output)

    # Store input as CSV
    content.to_csv(filename_output)

    # Cat output to command line
    with open(filename_output, 'r') as f:
        print(f.read())

    # Remove temporary file
    remove(filename_output) 

Then, you add (as explained in the other answers) the name of the python script to your .git/config and .gitattributes files.

.gitattributes:

*.xls diff=xls

.git/config:

[diff "xlsx"]
        textconv = python ./utils/xlsx_to_csv.py --input