Tracking Changes in Folder Structures

Problem

We are modifying files and folders direcly on the file server. With this practice there is a risk that changes in our folder structure, such as renaming, moving or deleting folders have undesired effects. For example, a link to a folder will be broken once the folder has been renamed. Or a script will not work any more because an input file or folder has been renamed or moved. Therefore, it is important to know what changes have been made to files and folders on the server.

Permanent Tracking Is Not an Option

It would be possible to let the operating system track such changes. However, this would slow down the server so much that this is not an option.

Proposal of a Simple Solution

We propose the following “hand-made” and simple solution to track changes: Once per day (preferably during the night) we run a script. This script gets a full list of files and folders that are currently stored on the server. This list does not only contain the paths to the files and folders but also additional information such as file sizes, times of last access and owners of the files. All these information are written to a text file.

These text files allow to

quickly search for a file or folder by just searching the text file. Even though these files have a size of about 50 Megabytes, searching these files using a good text editor is much more efficient than searching the whole file server.
compare text files referring to different days. Using the tools provided by professional version control software can support the comparison.

Supporting Functions

This package contains a set of functions that aim at supporting the approach described above.

Read and Save File Path Information

read_paths(): Read paths from a text file that contains exactly one path per line.

Recreate Folder Structure From Paths

write_paths_to_folder_tree(): Create a folder structure up to a certain folder depth in a given target directory. The folder structure is defined by a vector of paths. The idea of this function is to restore an old folder structure of which the paths have been saved. When two folder structures are restored from two path vectors that represent an original folder structrue at two different points in time, version control software can be used to compare these folder structures. In Tortoise SVN, for example, changes are marked with a red exclamation mark.
build_folders_from_file(): Restore a folder structure of which the paths were stored in a text file. This function first reads a vector of paths from a text file and then calls write_paths_to_folder_tree() to restore the so defined folder structure.