We Keep Coding

Copy only new or modified files/directories in C#

I am trying to create a simple “directory/file copy" console application in C#. What I need is to copy all folders and files (keeping the original hierarchy) from one drive to another, like from drive C:\Data to drive E:\Data.
However, I only want it to copy any NEW or MODIFIED files from the source to the destination.
If the file on the destination drive is newer than the one on the source drive, then it does not copy.
(the problem)
In the code I have, it's comparing file "abc.pdf" in the source with file "xyz.pdf" in the destination and thus is overwriting the destination file with whatever is in the source even though the destination file is newer. I am trying to figure out how to make it compare "abc.pdf" in the source to "abc.pdf" in the destination.
This works if I drill the source and destination down to a specific file, but when I back out to the folder level, it overwrites the destination file with the source file, even though the destination file is newer.
(my solutions – that didn’t work)
I thought by putting the “if (file.LastWriteTime > destination.LastWriteTime)” after the “foreach” command, that it would compare the files in the two folders, File1 source to File1 destination, but it’s not.
It seems I’m missing something in either the “FileInfo[]”, “foreach” or “if” statements to make this a one-to-one comparison. I think maybe some reference to the “Path.Combine” statement or a “SearchOption.AllDirectories”, but I’m not sure.
Any suggestions?
As you can see from my basic code sample, I'm new to C# so please put your answer in simple terms.
Thank you.
Here is the code I have tried, but it’s not working.
class Copy
{
public static void CopyDirectory(DirectoryInfo source, DirectoryInfo destination)
{
if (!destination.Exists)
{
destination.Create();
}
// Copy files.
FileInfo[] files = source.GetFiles();
FileInfo[] destFiles = destination.GetFiles();
foreach (FileInfo file in files)
foreach (FileInfo fileD in destFiles)
// Copy only modified files
if (file.LastWriteTime > fileD.LastWriteTime)
{
file.CopyTo(Path.Combine(destination.FullName,
file.Name), true);
}
// Copy all new files
else
if (!fileD.Exists)
{
file.CopyTo(Path.Combine(destination.FullName, file.Name), true);
}
// Process subdirectories.
DirectoryInfo[] dirs = source.GetDirectories();
foreach (DirectoryInfo dir in dirs)
{
// Get destination directory.
string destinationDir = Path.Combine(destination.FullName, dir.Name);
// Call CopyDirectory() recursively.
CopyDirectory(dir, new DirectoryInfo(destinationDir));
}
}
}

You can just take the array of files in "source" and check for a matching name in "destination"
/// <summary>
/// checks whether the target file needs an update (if it doesn't exist: it needs one)
/// </summary>
public static bool NeedsUpdate(FileInfo localFile, DirectoryInfo localDir, DirectoryInfo backUpDir)
{
bool needsUpdate = false;
if (!File.Exists(Path.Combine(backUpDir.FullName, localFile.Name)))
{
needsUpdate = true;
}
else
{
FileInfo backUpFile = new FileInfo(Path.Combine(backUpDir.FullName, localFile.Name));
DateTime lastBackUp = backUpFile.LastWriteTimeUtc;
DateTime lastChange = localFile.LastWriteTimeUtc;
if (lastChange != lastBackUp)
{
needsUpdate = true;
}
else
{/*no change*/}
}
return needsUpdate;
}

Update:
I modified my code with the suggestions above and all went well. It did exactly as I expected.
However, the problem I ran into was the amount of time it took run the application on a large folder. (containing 6,000 files and 5 sub-folders)
On a small folder, (28 files in 5 sub-folders) it only took a few seconds to run. But, on the larger folder it took 35 minutes to process only 1,300 files.
Solution:
The code below will do the same thing but much faster. This new version processed 6,000 files in about 10 seconds. It processed 40,000 files in about 1 minute and 50 seconds.
What this new code does (and doesn’t do)
If the destination folder is empty, copy all from the source to the destination.
If the destination has some or all of the same files / folders as the source, compare and copy any new or modified files from the source to the destination.
If the destination file is newer than the source, don’t copy.
So, here’s the code to make it happen. Enjoy and share.
Thanks to everyone who helped me get a better understanding of this.
using System;
using System.IO;
namespace VSU1vFileCopy
{
class Program
{
static void Main(string[] args)
{
const string Src_FOLDER = #"C:\Data";
const string Dest_FOLDER = #"E:\Data";
string[] originalFiles = Directory.GetFiles(Src_FOLDER, "*", SearchOption.AllDirectories);
Array.ForEach(originalFiles, (originalFileLocation) =>
{
FileInfo originalFile = new FileInfo(originalFileLocation);
FileInfo destFile = new FileInfo(originalFileLocation.Replace(Src_FOLDER, Dest_FOLDER));
if (destFile.Exists)
{
if (originalFile.Length > destFile.Length)
{
originalFile.CopyTo(destFile.FullName, true);
}
}
else
{
Directory.CreateDirectory(destFile.DirectoryName);
originalFile.CopyTo(destFile.FullName, false);
}
});
}
}
}

Get file names from Resources sub folder

In my Resources folder I have a subfolder for images, I would like to get all the file names of those images from within that folder.
tried several Resources.loadAll methods to afterwards get the .name but without success
was is the right practice to achieve what I'm trying to do here ?

There is no built-in API to do this because the information is not after you build. You cant' even do this with what's in the accepted answer. That would only work in the Editor. When you build the project, your code will fail.
Here's what to do:
1. Detect when the build button is clicked or when a build is about to happen in the OnPreprocessBuild function.
2. Get all the file names with Directory.GetFiles, serialize it to json and save it to the Resources folder. We use json to make it easier to read individual file name. You don't have to use json. You must exclude the ".meta" extension.
Step 1 and 2 are done in the Editor.
3. After a build or during run-time, you can access the saved file that contains the file names as a TextAsset with Resources.Load<TextAsset>("FileNames") then de-serialize the json from TextAsset.text.
Below is very simplified example. No error handling and that's up to you to implement. The Editor script below saves the file names when you click on the Build button:
[Serializable]
public class FileNameInfo
{
public string[] fileNames;
public FileNameInfo(string[] fileNames)
{
this.fileNames = fileNames;
}
}
class PreBuildFileNamesSaver : IPreprocessBuildWithReport
{
public int callbackOrder { get { return 0; } }
public void OnPreprocessBuild(UnityEditor.Build.Reporting.BuildReport report)
{
//The Resources folder path
string resourcsPath = Application.dataPath + "/Resources";
//Get file names except the ".meta" extension
string[] fileNames = Directory.GetFiles(resourcsPath)
.Where(x => Path.GetExtension(x) != ".meta").ToArray();
//Convert the Names to Json to make it easier to access when reading it
FileNameInfo fileInfo = new FileNameInfo(fileNames);
string fileInfoJson = JsonUtility.ToJson(fileInfo);
//Save the json to the Resources folder as "FileNames.txt"
File.WriteAllText(Application.dataPath + "/Resources/FileNames.txt", fileInfoJson);
AssetDatabase.Refresh();
}
}
During run-time, you can retrieve the saved file names with the example below:
//Load as TextAsset
TextAsset fileNamesAsset = Resources.Load<TextAsset>("FileNames");
//De-serialize it
FileNameInfo fileInfoLoaded = JsonUtility.FromJson<FileNameInfo>(fileNamesAsset.text);
//Use data?
foreach (string fName in fileInfoLoaded.fileNames)
{
Debug.Log(fName);
}

Hmm... why not try this.
using System.IO;
Const String path = ""; /file path
private void GetFiles()
{
string [] files = Directory.GetFiles (path, "*.*");
foreach (string sourceFile in files)
{
string fileName = Path.GetFileName (sourceFile);
Debug.Log("fileName");
}
}

Application: Application Launcher, can't Move directory, it's being used by another process

I'm writing application launcher as a Window Application in C#, VS 2017. Currently, having problem with this piece of code:
if (System.IO.Directory.Exists(extractPath))
{
string[] files = System.IO.Directory.GetFiles(extractPath);
string[] dirs = Directory.GetDirectories(extractPath);
// Copy the files and overwrite destination files if they already exist.
foreach (string s in files)
{
// Use static Path methods to extract only the file name from the path.
var fileName = System.IO.Path.GetFileName(s);
var destFile = System.IO.Path.Combine(oldPath, fileName);
System.IO.File.Move(s, destFile);
}
foreach (string dir in dirs)
{
//var dirSplit = dir.Split('\\');
//var last = dirSplit.Last();
//if (last != "Resources")
//{
var fileName = System.IO.Path.GetFileName(dir);
var destFile = System.IO.Path.Combine(oldPath, fileName);
System.IO.Directory.Move(dir, destFile);
//}
}
}
I'm getting well known error
"The process cannot access the file 'XXX' because it is being used by another process."
I was looking for solution to fix it, found several on MSDN and StackOvervflow, but my problem is quite specific. I cannot move only 1 directory to another, which is Resources folder of my main application:
Here is my explanation why problem is specific:
I'm not having any issues with moving other files from parent directory. Error occurs only when loop reaches /Resources directory.
At first, I was thinking that it's beeing used by VS instantion, in which I've had main app opened. Nothing have changed after closing VS and killing process.
I've copied and moved whole project to another directory. Never opened it in VS nor started via *.exe file, to make sure that none of files in new, copied directory, is used by any process.
Finally, I've restarted PC.
I know that this error is pretty common when you try to Del/Move files, but in my case, I'm sure that it's being used only by my launcher app. Here is a little longer sample code to show what files operation I'm actually doing:
private void RozpakujRepo()
{
string oldPath = #"path\Debug Kopia\Old";
string extractPath = #"path\Debug Kopia";
var tempPath = #"path\ZipRepo\RexTempRepo.zip";
if (System.IO.File.Exists(tempPath) == true)
{
System.IO.File.Delete(tempPath);
}
System.IO.Compression.ZipFile.CreateFromDirectory(extractPath, tempPath);
if (System.IO.Directory.Exists(oldPath))
{
DeleteDirectory(oldPath);
}
if (!System.IO.Directory.Exists(oldPath))
{
System.IO.Directory.CreateDirectory(oldPath);
}
if (System.IO.Directory.Exists(extractPath))
{
string[] files = System.IO.Directory.GetFiles(extractPath);
string[] dirs = Directory.GetDirectories(extractPath);
// Copy the files and overwrite destination files if they already exist.
foreach (string s in files)
{
// Use static Path methods to extract only the file name from the path.
var fileName = System.IO.Path.GetFileName(s);
var destFile = System.IO.Path.Combine(oldPath, fileName);
System.IO.File.Move(s, destFile);
}
foreach (string dir in dirs)
{
//var dirSplit = dir.Split('\\');
//var last = dirSplit.Last();
//if (last != "Resources")
//{
var fileName = System.IO.Path.GetFileName(dir);
var destFile = System.IO.Path.Combine(oldPath, fileName);
System.IO.Directory.Move(dir, destFile);
//}
}
}
string zipPath = #"path\ZipRepo\RexRepo.zip";
ZipFile.ExtractToDirectory(zipPath, extractPath);
}
And now, my questions:
Can it be related to file types (.png, .ico, .bmp) ?
Can it be related to fact, that those resources files are being used like, as, for example .exe file icon in my main application? Or just because those are resources files?
Is there anything else what I'm missing and what can cause the error?
EDIT:
To clarify:
There are 2 apps:
Main Application
Launcher Application (to launch Main Application)
And Resources folder is Main Application/Resources, I'm moving it while I'm doing application version update.

It appeared that problem is in different place than in /Resources directory. Actually problem was with /Old directory, because it caused inifinite recurrence.

Get path to executable from command (as cmd does)

Given a command-line style path to a command such as bin/server.exe or ping, how can I get the full path to this executable (as cmd or Process.Start would resolve it)?
I tried Path.GetFullPath, but it always expands relative to the working directory. It expands bin/server.exe correctly, however given ping it returns c:\users\matt\ping (non-existent). I want c:\Windows\system32\ping.exe.
Edit: I would like the same behaviour as cmd. Some considerations:
When there is a local executable with the same name as one in the path, cmd prefers the local one
cmd can expand the command server to server.bat or server.exe (adding the file extension)
I also tried Windows' command-line tool called where . It does almost I want:
Displays the location of files that match the search pattern. By default, the search is done along the current directory and in the paths specified by the PATH environment variable.
>where ping
C:\Windows\System32\PING.EXE
>where bin\server
INFO: Could not find files for the given pattern(s).
(This question is hard to search around because of the two different meanings of the word 'path')

Considering PATHEXT too, stealing from Serj-Tm's answer (sorry! +1 to him):
public static string WhereSearch(string filename)
{
var paths = new[]{ Environment.CurrentDirectory }
.Concat(Environment.GetEnvironmentVariable("PATH").Split(';'));
var extensions = new[]{ String.Empty }
.Concat(Environment.GetEnvironmentVariable("PATHEXT").Split(';')
.Where(e => e.StartsWith(".")));
var combinations = paths.SelectMany(x => extensions,
(path, extension) => Path.Combine(path, filename + extension));
return combinations.FirstOrDefault(File.Exists);
}
Sorry the indentation's a bit all-over-the-place - I was trying to make it not scroll. I don't know if the StartsWith check is really necessary - I'm not sure how CMD copes with pathext entries without a leading dot.

public static string GetFullPath(string filename)
{
return new[]{Environment.CurrentDirectory}
.Concat(Environment.GetEnvironmentVariable("PATH").Split(';'))
.Select(dir => Path.Combine(dir, filename))
.FirstOrDefault(path => File.Exists(path));
}

If you're only interested in searching the current directory and the paths specified in the PATH environment variable, you can use this snippet:
public static string GetFullPath(string fileName)
{
if (File.Exists(fileName))
return Path.GetFullPath(fileName);
var values = Environment.GetEnvironmentVariable("PATH");
foreach (var path in values.Split(';'))
{
var fullPath = Path.Combine(path, fileName);
if (File.Exists(fullPath))
return fullPath;
}
return null;
}

You have to search the entire disk.
Windows can respond to things like, iexplore, ping, cmd, etc, because they are in the registry under this key:
HKEY_LOCAL_MACHINE
SOFTWARE
Microsoft
Windows
CurrentVersion
App Paths
The only other way is to search the entire disk for the application.
EDIT: My understanding was, that you want to search for any random executable name, not the ones that are already known to Windows..

internal class Program
{
static void Main(string[] args)
{
string fullPath = GetExactPathFromEnvironmentVar("ping.exe");
if (!string.IsNullOrWhiteSpace(fullPath))
Console.WriteLine(fullPath);
else
Console.WriteLine("Not found");
}
static string GetExactPathFromEnvironmentVar(string program)
{
var pathVar = System.Environment.GetEnvironmentVariable("PATH");
string[] folders = pathVar.Split(';');
foreach (var folder in folders)
{
string path = Path.Combine(folder, program);
if (File.Exists(path))
{
return path;
}
}
return null;
}
}
HTH

How do I compare one collection of files to another in c#?

I am just learning C# (have been fiddling with it for about 2 days now) and I've decided that, for leaning purposes, I will rebuild an old app I made in VB6 for syncing files (generally across a network).
When I wrote the code in VB 6, it worked approximately like this:
Create a Scripting.FileSystemObject
Create directory objects for the source and destination
Create file listing objects for the source and destination
Iterate through the source object, and check to see if it exists in the destination
if not, create it
if so, check to see if the source version is newer/larger, and if so, overwrite the other
So far, this is what I have:
private bool syncFiles(string sourcePath, string destPath) {
DirectoryInfo source = new DirectoryInfo(sourcePath);
DirectoryInfo dest = new DirectoryInfo(destPath);
if (!source.Exists) {
LogLine("Source Folder Not Found!");
return false;
}
if (!dest.Exists) {
LogLine("Destination Folder Not Found!");
return false;
}
FileInfo[] sourceFiles = source.GetFiles();
FileInfo[] destFiles = dest.GetFiles();
foreach (FileInfo file in sourceFiles) {
// check exists on file
}
if (optRecursive.Checked) {
foreach (DirectoryInfo subDir in source.GetDirectories()) {
// create-if-not-exists destination subdirectory
syncFiles(sourcePath + subDir.Name, destPath + subDir.Name);
}
}
return true;
}
I have read examples that seem to advocate using the FileInfo or DirectoryInfo objects to do checks with the "Exists" property, but I am specifically looking for a way to search an existing collection/list of files, and not live checks to the file system for each file, since I will be doing so across the network and constantly going back to a multi-thousand-file directory is slow slow slow.
Thanks in Advance.

The GetFiles() method will only get you files that does exist. It doesn't make up random files that doesn't exist. So all you have to do is to check if it exists in the other list.
Something in the lines of this could work:
var sourceFiles = source.GetFiles();
var destFiles = dest.GetFiles();
foreach (var file in sourceFiles)
{
if(!destFiles.Any(x => x.Name == file.Name))
{
// Do whatever
}
}
Note: You have of course no guarantee that something hasn't changed after you have done the calls to GetFiles(). For example, a file could have been deleted or renamed if you try to copy it later.
Could perhaps be done nicer somehow by using the Except method or something similar. For example something like this:
var sourceFiles = source.GetFiles();
var destFiles = dest.GetFiles();
var sourceFilesMissingInDestination = sourceFiles.Except(destFiles, new FileNameComparer());
foreach (var file in sourceFilesMissingInDestination)
{
// Do whatever
}
Where the FileNameComparer is implemented like so:
public class FileNameComparer : IEqualityComparer<FileInfo>
{
public bool Equals(FileInfo x, FileInfo y)
{
return Equals(x.Name, y.Name);
}
public int GetHashCode(FileInfo obj)
{
return obj.Name.GetHashCode();
}
}
Untested though :p

One little detail, instead of
sourcePath + subDir.Name
I would use
System.IO.Path.Combine(sourcePath, subDir.Name)
Path does reliable, OS independent operations on file- and foldernames.
Also I notice optRecursive.Checked popping out of nowhere. As a matter of good design, make that a parameter:
bool syncFiles(string sourcePath, string destPath, bool checkRecursive)
And since you mention it may be used for large numbers of files, keep an eye out for .NET 4, it has an IEnumerable replacement for GetFiles() that will let you process this in a streaming fashion.