unzipfile fails to extract files larger than 4 GiB

[Diogo Pereira]

I'm trying to use unzipfile in Stata 14.1 to extract a ~1 GiB zip file containing a ~5 GiB text file. unzipfile fails with

I/O error
could not perform unzip
r(601)

It seems to extract the first 2³⁰ - 1 bytes correctly.

The file system is NTFS and I can extract the file with other tools (7-zip and the built-in Windows tool).

[Diogo Pereira]

This still does not work in Stata 15.

[William Lisowski]

Apparently Stata is not among those who support the ZIP64 format extensions as described here in Wikipedia.

[Clyde Schechter]

If the reason for doing this is to include unzipping the file as part of a running script in a do-file, and you have other programs that can do the unzipping, you can still accomplish that by using the -!windows_command- syntax which forks a Windows shell process that executes the Windows command, waits for it to complete, and then returns control to the do-file. And, you can even reference Stata local macros inside the windows_command: Stata dereferences the macros and passes the command to Windows with the values of the macros substituted in.

[Kevin Crow (StataCorp)]

The file size limit is a result of the library used to handle the back-end [de]compression routines. Switching the code to use a 64-bit library will fix this issue. I will get the fix out in a future update to Stata.

[Kevin Crow (StataCorp)]

This issue has been fixed in the 15.1 update.

[Diogo Pereira]

This issue has been fixed in the 15.1 update.

Thanks. I have just updated and verified that it now works as expected.