A CRC error ("CRC failed", "checksum error" or "data error") means a file inside the archive no longer matches the checksum recorded when the archive was made, so its data has been changed or damaged. "Unexpected end of archive" means the archive stops before it should, almost always because a download or copy was cut short. The fastest fix for both is a fresh, complete copy from the source. If there is no other copy, an archive repair tool can usually save every undamaged file and recover the damaged ones as far as possible.
Below is what each error really means, how to tell which problem you have, and what to do about it on Windows 10 and 11.
What a CRC error actually means
CRC stands for cyclic redundancy check. When a ZIP or RAR archive is created, the archiver calculates a CRC-32 value for every file it adds: a short number that acts like a fingerprint of that file's contents. It stores the fingerprint next to the compressed data.
When you extract the archive, the extractor decompresses each file, calculates the fingerprint again and compares the two. If they differ, even by a single bit, you get a CRC error. The archive is telling you, accurately, that the file you are about to get is not the file that went in.
A few things worth knowing:
- A CRC error is usually per file. One damaged photo inside a ZIP of 500 photos does not mean the other 499 are bad.
- Many extractors stop or delete the output when they hit a CRC error, so you may get nothing, even when the damage is tiny.
- In compressed data, one damaged byte can garble the rest of that file from that point on, which is why the affected file is often only partly recoverable.
- A CRC error on a password-protected archive can also mean the wrong password was entered, because wrong-password decryption produces data that fails the check.
What "unexpected end of archive" means
This error means the extractor reached the physical end of the file while it was still expecting more data. The archive has been truncated.
In a ZIP file, that is especially disruptive. ZIPs keep their table of contents, the central directory, at the very end. Cut off the end and the table of contents goes with it, which is why Windows File Explorer often reports the whole compressed folder as invalid. The good news: the files themselves are stored earlier in the archive, so everything before the cut-off point is usually still there and recoverable.
Common causes of truncation:
- A browser download that failed or was cancelled (look for leftover
.crdownloador.partfiles). - A copy to a USB stick or memory card that was unplugged too early or not ejected safely.
- An archive larger than 4 GB copied to a FAT32-formatted drive, which cannot hold files that big.
- The PC crashing or losing power while the archive was being created.
- A file recovered from a failed drive where only part of it could be read.
How to tell which problem you have
| What you see | Likely cause | Best first step |
|---|---|---|
| "Unexpected end of archive", file smaller than expected | Incomplete download or copy | Download or copy it again and compare sizes |
| CRC error on one or a few files, rest extract fine | Damaged bytes in the middle of the archive | Get a fresh copy; otherwise repair to save the rest |
| CRC error on every file of a protected archive | Wrong password | Re-enter the password carefully (check Caps Lock and keyboard layout) |
| CRC errors that change each time you copy it | Failing drive, cable, USB port or memory | Copy to a different drive; check the hardware |
| Both errors together | Truncated archive with damage near the end | Repair to recover everything before the damage |
Fixes to try first
- Re-download or re-copy the archive. This solves the large majority of cases. If the site publishes a file size or checksum, compare it. In PowerShell,
Get-FileHash .\archive.zipshows the SHA-256 hash for comparison. - Ask the sender to send it another way. Email systems and some chat apps limit attachment sizes; a cloud share link is often more reliable for large archives.
- Copy it to your internal drive. If the archive is on removable media or a network share, copy it locally before extracting. If copying fails or produces different results each time, suspect the drive or connection.
- Look for an earlier version. Right-click the file and choose Properties > Previous Versions, or use Version history in OneDrive.
- Try another extractor on a copy. Some tools let you keep files that fail the CRC check instead of deleting them. That can be enough for a text file, but a partly broken photo, video or document still needs repairing.
If your RAR archive was created with a recovery record, the archiver that made it may be able to fix small amounts of damage using that record. It is worth checking before anything else. See how to repair a corrupted RAR file.
Repairing the archive when there is no other copy
When the archive you have is the only one, File Repairer can fix CRC errors and "unexpected end of archive" problems in both ZIP and RAR files. It works through the archive file by file instead of relying on the damaged table of contents:
- Intact files are kept exactly as they were. If a file passes its CRC check, it is saved byte for byte.
- Damaged files are recovered as far as possible and marked "(partial)" in their names, so you know which ones to double-check.
- You see the result before saving. The preview lists the archive's contents with a "what was recovered" panel.
- Your original is untouched. The repaired archive is saved as a new file with " (repaired)" in its name. RAR contents are saved to a folder and to a new ZIP.
To do it, drag the archive into File Repairer, run the repair, review the preview and save. Checking, repairing and previewing are free; saving needs an Installation Key. You can also drag in a whole folder of archives and use Repair All. The step-by-step walkthrough is in our guide to repairing a corrupted ZIP file.
A partial file from inside the archive may itself be repairable. If a Word document, PDF, photo or video comes out marked "(partial)", add it to File Repairer as a separate file and repair it on its own.
What can't be fixed, and how to avoid it next time
Repair recovers what is physically in the file. If an archive was truncated, the data after the cut-off point was never written, so files stored there cannot be brought back. Likewise, bytes that were overwritten are gone. File Repairer also cannot recover deleted archives or read failing drives; if Windows cannot copy the file at all, use data recovery software first and then repair what it recovers.
To avoid these errors in future:
- Let downloads finish completely and check the size before deleting the source.
- Always use Safely Remove Hardware and Eject Media before unplugging USB drives and memory cards.
- Format external drives as exFAT or NTFS if you work with archives over 4 GB.
- Test important archives after creating them, before you delete the originals.
More habits like these are in our guide to preventing file corruption.
Frequently asked questions
Does a CRC error mean the whole archive is ruined?
Usually not. The check is done per file, so a CRC error normally points to one or a few damaged files while the others are fine.
Can a wrong password cause a CRC error?
Yes. With an encrypted archive, a mistyped password produces scrambled output that fails the checksum. If every file fails, try the password again before assuming damage.
Can I get back the part of the archive that was cut off?
No. Data after the point where the download or copy stopped was never saved, so only files stored before that point can be recovered.
