How to extract all files from a subdirectory in an archive?

[gbook]

If I have an archive with paths, such as:

data
data/1
data/1/2

How would I extract all the files stored in data/1/2 into a directory on disk?

[rikyoz]

Hi!
If you're using BitFileExtractor, you can use the extractMatching method, which takes a second parameter with a wildcard filter:

extractor.extractMatching( "path/to/archive.7z", "/data/1/2/*", "out" );

Unfortunately, there's no equivalent method in BitArchiveReader, as I'm still working on "synchronizing" the capabilities of both classes; hopefully I'll add such method in v4.1.

At the moment, with BitArchiveReader you can extract specific items using a vector of indices, so you could write something like this to achieve the same result:

std::vector< std::uint32_t > matchingIndices;
for ( const auto& item : reader ) {
    if ( item.path().rfind( "data/1/2/", 0 ) != 0 ) { // Or C++20's starts_with method.
       matchingIndices.push_back( item.index() );
    }
}
reader.extractTo( "out", matchingIndices );

[gbook]

I tried the BitFileExtractor method

extractor.extractMatching("pathTo.7z", "/data/1/2/*", "C:/temp/data/1/2");

If I use 7zip to open the archive, I see that there are files in that path within the archive, but nothing is extracted to disk when I run that function. Is there a return value for this function, or a way to tell if it successfully matched any files in the archive?

[rikyoz]

Sorry, I forgot a few details in my previous comment.

First, 7-Zip usually reports the paths of items inside archives without the / prefix. There are some edge cases (e.g. if the / is explicitly specified for the item name when creating the archive), but this is unlikely to be your case.

Second, 7-Zip reports the paths of items inside archives using the OS native path separator, which is \ on Windows.

So on Windows, this should be the correct code:
extractor.extractMatching("pathTo.7z", "data\\1\\2\\*", "C:/temp/data/1/2");

If I use 7zip to open the archive, I see that there are files in that path within the archive, but nothing is extracted to disk when I run that function. Is there a return value for this function, or a way to tell if it successfully matched any files in the archive?

This is strange, as the function should throw an exception if no match is found in the archive.

[gbook]

That worked! Thank you

…

On Mon, Jan 13, 2025 at 3:39 PM Riccardo ***@***.***> wrote: Sorry, I forgot a few details in my previous comment. First, 7-Zip *usually* reports the paths of items inside archives without the / prefix. There are some edge cases (e.g. if the / is explicitly specified for the item name when creating the archive), but this is unlikely to be your case. Second, 7-Zip reports the paths of items inside archives using the OS native path separator, which is \ on Windows. So on Windows, this should be the correct code: extractor.extractMatching("pathTo.7z", "data\\1\\2\\*", "C:/temp/data/1/2"); If I use 7zip to open the archive, I see that there are files in that path within the archive, but nothing is extracted to disk when I run that function. Is there a return value for this function, or a way to tell if it successfully matched any files in the archive? This is strange, as the function should throw an exception if no match is found in the archive. — Reply to this email directly, view it on GitHub <#270 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB7K5B7HPNCJFPEVVWPV4ZD2KQP77AVCNFSM6AAAAABU4THTXCVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCOBSGQ2TGOA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[gbook]

One more question... I want to extract the files located in data/1/2/* into the directory C:/temp/data/1/2/

What I actually get is C:/temp/data/1/2/data/1/2/

Is there a way to ignore the internal archive path when extracting the files to disk?

[rikyoz]

Is there a way to ignore the internal archive path when extracting the files to disk?

If you only have files inside the data/1/2/ path, you can use the extractor.setRetainDirectories(false) method.
However, if you also have subdirectories inside data/1/2/, the setRetainDirectories(false) would drop them too (data/1/2/3/file.txt becomes just file.txt, so when you extract to C:\temp\data\1\2, you will have C:\temp\data\1\2\file.txt instead of C:\temp\data\1\2\3\file.txt).

To be honest, I am a little embarrassed at how little flexibility the API provided by bit7z has in use cases like this. I will absolutely work to improve it.
In the upcoming v4.1 there is already some improvement in this regard, as it introduces RenameCallbacks, which give the user the ability to customize the name/path of the extracted files at will.
I hope to implement further improvements to the API to make it more user-friendly.

[gbook]

I want to revisit this topic of extracting sub-directories :-)

Imagine I have a path in the 7zip archive that I want to extract to disk, but I do want to retain the subdirectory structure. For example I have

data/1/file.txt
data/1/2/file.txt
data/1/2/3/file.txt

Is it possible to extract data/1/* to C:\temp\extracted\ , so that it appears on disk like this?

C:\temp\extracted\file.txt
C:\temp\extracted\2\file.txt
C:\temp\extracted\2\3\file.txt

[rikyoz]

At the moment, no, unfortunately.
But since you opened this discussion, I started working on a new method in the BitInputArchive class that will allow exactly your use case.
My idea is to add a method as follow:

void extractFolderTo( const tstring& outDir,
                      const tstring& folderPath,
                      FolderPathPolicy policy = FolderPathPolicy::Strip ) const;

where FolderPathPolicy is an enumeration with the following possible values:

enum class FolderPathPolicy : std::uint8_t {
    Strip,    ///< Remove the folder path from the extracted path.
    KeepName, ///< Preserve the folder name in the extracted path.
    KeepPath, ///< Preserve the full folder path in the extracted path.
};

In your case, then, you would write something like this:

// The default behavior, i.e. your use case -> C:\temp\extracted\file.txt ...
reader.extractFolderTo("C:\\temp\\extracted", "data\\1"); 
// Other possible cases:
//
// Keep the name of the folder (i.e., "1") -> C:\temp\extracted\1\file.txt ...
// reader.extractFolderTo("C:\\temp\\extracted", "data\\1", FolderPathPolicy::KeepName)
//
// Keep the whole path of the folder (i.e., "data\\1") -> C:\temp\extracted\data\1\file.txt ...
// reader.extractFolderTo("C:\\temp\\extracted", "data\\1", FolderPathPolicy::KeepPath)

I still need to push the changes to develop, as I'm still working on and polishing the implementation of the method; I hope to include it either in the v4.0.10 or in the v4.1-beta.

[gbook]

That looks great. I look forward to the addition!