UUnpackNest
EN
EnglishالعربيةEspañolPortuguêsFrançaisDeutsch
Selective extract

How to Extract One File Without Unpacking the Whole Archive

A 3 GB archive arrived and you need a 200 KB PDF that is somewhere inside it. Unpacking everything is how phones run out of space and how people give up and delete the download. Every UnpackNest tool can write a single entry. Whether that is also quick depends on the format, not on the button.

Updated 2026-10-22 · Written and tested by the UnpackNest editorial team (TECCONET)

List first, on purpose

You cannot pick a file you have not seen. Listing is cheap for ZIP, because the directory is a table at the end. It is a walk for TAR. It may require a password for a 7z whose names are encrypted. Do the list on the matching tool, read the names, and only then extract.

  1. ZIP: open the ZIP tool, wait for the entry count, press Extract on one row.
  2. RAR: open the RAR tool, leave encrypted and blocked rows alone, tick one safe file, press Extract selected. Extract all safe files is the button you are trying not to press.
  3. 7z: open the 7z tool. If a lock is on that row only, type the password when asked and extract that row. You do not have to unlock every neighbour if the headers were visible without a password.
  4. TAR or gzip: open the TAR tool and extract the one path you need. A lone .gz that is a single CSV has only one member. There is nothing to be selective about.

Save into a folder you can find, ideally not the same folder the archive came from if that folder is a crowded Downloads. The saved name is the inner file name. If two archives both contain readme.txt, the second save may become readme (1).txt. Open the one with today's time.

ZIP is the format this is easy for

Each ZIP entry is its own compressed stream, and the central directory stores where that stream starts. The tool reads the directory, seeks to the entry, checks the CRC-32, and saves. The other gigabytes stay compressed and untouched. This is the right default when you are on a phone, on a Chromebook with a small disk, or on a metered connection where you already paid to download the ZIP and do not want a second copy of every photo.

You still downloaded the whole ZIP. Selective extract does not fetch one member from a server. Nobody can pull chapter3.pdf out of a ZIP URL without the rest of the file, because the directory that points at chapter 3 is at the end. If the download was cut off before the end, you may have neither the directory nor the file. That is truncation, handled in the repair guide, not a reason to extract at random.

Folders in the name (scans/2024/form.pdf) are part of the entry name. You get the PDF, not a recreated tree, from the browser save dialog. If you need the tree, extract on a desktop with "full pathnames" enabled, or create the folders yourself and move the file. For one form, the flat save is a feature.

RAR depends on the sender's settings

A non-solid RAR stores each file so a reader can reach it. Tick one row and extract. A solid RAR shares a compression block. The same click still saves only your file, but the engine may decode a long prefix first. If the status line sits on that one file for a surprisingly long time, read the solid-compression guide before you cancel and start over. Cancelling and retrying repeats the prefix.

Blocked rows cannot be the "one file". An encrypted entry, a link, or an unsafe path will not be written because you selected only that row. Choose a different entry or take the archive to a desktop reader if the encrypted file is truly the one you need and you have the password. The browser RAR tool will not ask for the password.

Multi-part RAR cannot be mined for one file until the volumes are joined on a computer. There is no partial credit for owning part1.

7z is often solid, so budget time

The 7z page will extract one index from the list. Budget for the solid block. A 1 GB per-file cap still applies to that one entry: if the PDF itself expands past the cap, the tool refuses it even though you asked for nothing else. That refusal is about the output size, not about the rest of the archive.

Encrypted file names change the order. You type the password to see the list, then you extract one row. The password is held in the tab for that operation. Clear it by reloading the page when you are done, on a shared computer. Reloading discards the archive too. It was never uploaded.

If you need twenty files and not one, select them in a desktop 7-Zip window rather than clicking twenty times in a browser. The browser flow is aimed at the "one contract inside a dump" case. It is a poor batch processor.

TAR and the "I only need the license file" case

Source tarballs hide the file you want under a top-level directory: project-1.4/LICENSE. The TAR tool shows that path. Extract that entry. You do not need the rest of the tree to read a license or a single CSV. If the path is longer than you can see, widen the window or copy the name from the row before you forget which index.js you meant. Several packages ship more than one.

A .tar.xz is outside the browser TAR tool. On a computer, tar -xOf archive.tar.xz project-1.4/LICENSE > LICENSE writes just that member. The -O flag sends it to the terminal or to the redirect you chose. Listing first with tar -tf stops you extracting the wrong LICENSE from a bundled third-party folder.

Disk space, twice

People forget that the archive and the extracted file both occupy space, and that a failed extract can leave a partial download beside them. Before you start on a phone:

"Extract all" on a 3 GB archive of photos can ask for 3 GB more. That is the operation this page is helping you avoid.

Checking you got the right file

Open the saved file in its real application. A PDF that does not open, or a spreadsheet that warns about corruption, is a bad extract or a bad entry, not a reason to unpack everything "in case". Go back to the tool and read whether that row reported a CRC or checksum problem. If it did, the other rows may be fine and this one needs a new copy of the archive. If it did not, the file is intact and the application is complaining about the document itself (a scan that is not a spreadsheet, a CSV named .xlsx).

Then you can delete the archive, if you are sure you will not need a second file from it tomorrow. Keeping it for a day is cheaper than downloading it again on a metered connection.

Frequently asked questions

The list shows the file but Extract does nothing

Look for a blocked or encrypted flag on that row. The button will not override it. Pick a different row to confirm the tool is alive. If every row fails, the archive itself is the problem (truncation, wrong password on a 7z header, unsupported method).

I extracted one file and the folder structure is gone

The browser saved the file, not the directories in its path. The path is still visible in the list if you need to know where it lived. Desktop 7-Zip's extract dialog has a path option if you need the folders recreated. For a single form, you do not.

Will extracting one file be faster on ZIP than on 7z?

Almost always, yes, when the 7z is solid and the ZIP is Deflate. The button count is the same. The CPU work is not. If you create archives for other people and they only ever need one file, send a non-solid ZIP.

Can I extract one file from a link without downloading the archive?

No. The tool reads a local file. "Online" here means the page loads from the web and the archive stays on the device. It does not mean the server plucks one member out of your cloud link. Download the archive, then extract one entry.

Open and extract ZIP files in your browserAll guides

Tools for this task

Related guides