Buffering the output of a "dir" command

[swiftie]

I had a command prompt open in a directory that contained 8 files. I wanted to keep an inventory of the 8 files, so I ran the command:

C:\>dir > inventory.2015-06-03

When I looked at the contents of inventory.2015-06-03 I was surprised to find that it contained the inventory.2015-06-03 file itself, as a 0 byte file.

Is there a technique to buffer the output from the "dir" command before the ">" takes effect, so I don't get the extra file in my listing?

 

[GokAy]

You can try:
dir | findstr /v /i "\myfile.txt$" > myfile.txt

in case symbols don't make it

 

[swiftie]

Thanks - your example works perfectly (of course).

It's interesting to find the use of Regular Expressions in a Windows context. For a moment I thought I'd strayed into a Linux discussion.

 

[GokAy]

I don't know much about switches and whatnot, searched google with: cmd dir exclude files

this came up:
Excluding files of particular extension using DIR command on windows command line - Super User

then added your "> text.file" part to it, tried in my PC, posted for you

 

[Ztruker]

No regular expression there: dir | findstr /v /i "\myfile.txt$" > myfile.txt

In english this says:

Display the list of files and folders in the current directory.
Pipe the output to the find string command. Make it case insensitive and exclude the file myfile.txt$ ($ meaning anything)
then redirect the output to myfile.txt.

 

[Pyprohly]

No regular expression there: dir | findstr /v /i "\myfile.txt$" > myfile.txt

The text in red is a regular expression. At least it's supposed to be.

GokAy, your regex, \myfile.txt$, is slightly incorrect.


Recalling a few special character meanings in regular expressions:

• "\" is an escape character,
• "." represents any single character,
• "$" at the end of a regex means that the string to match should only be matched at the end of a line (immediately before any line break).

So the regex \myfile.txt$ actually says to match the string literal myfile, any character, and the string literal txt. $ at the end of a regex says that the string to match must be at the end of a line, before a line break character. Escaping the literal character m doesn't do anything.

This particular regular expression can be expressed in wildcard notation too. Its wildcard notation would be myfile?txt (where ? in wildcard notation represents any single character). This, however, will not say to match at the end of a line.

The correct regular expression that should be used in the Findstr command should be myfile\.txt$.


I presume you've already realised this by now, but there is no way to extend the output buffer of commands, Swiftie. GokAy's solution of filtering the output of Dir is exactly how your problem should be tackled.

 

[swiftie]

Thanks to all for the help offered. It was the "$" in "\myfile.txt$" that showed me it was a regular expression. I hadn't noticed that the "\" and the "." were not quite right, but I no longer use regular expressions much since I retired (from IBM)