PS_09_Perl - Files and I/O

Files and I/O

Filehandles:

A filehandle is the name for the connection between your Perl program and the operating system. Filehandles follow the same naming conventions as labels, and occupy their own namespace.

Every Perl program has three filehandles that are automatically opened for it: STDIN,

STDOUT, and STDERR:

STDIN Standard input (keyboard or file)

STDOUT Standard output (print and write send output here)

STDERR Standard error (channel for diagnostic output)

Filehandles are created using the open() function:

open(FILE,”filename”);

You can open files for reading, writing, or appending:

open(FILE,”> newout.dat”) Writing, creating a new file

open(FILE,”>> oldout.dat”) Appending to existing file

open(FILE,”< input.dat”) Reading from existing file

As an aside, under Windows, there are a number of ways to refer to the full path to a file:

”c:\\temp\\file” Escape the backslash in double quotes

‘c:\temp\file’ Use proper path in single quotes

“c:/temp/file” UNIX-style forward slashes

It is important to realize that calls to the open() function are not always successful. Perl will not (necessarily) complain about using a filehandle created from a failed open().

This is why we test the condition of the open statement:

open(F,”< badfile.dat”) or die “open: $!”

You may wish to test for the existence of a file or for certain properties before opening it.

Fortunately, there are a number of file test operators available:

File test Meaning

-e file File or directory exists

-T file File is a text file

-w file File is writable

-r file File is readable

-s file File exists and has nonzero length

These operators are usually used in a conditional expression:

if (-e myfile.dat) {

open(FILE,”< myfile.dat”) or die “open: $!\n”;

}

Even more information can be obtained about a given file using the stat() function.

Using filehandles

After a file has been opened for reading you can read from it using the diamond operator just as you have already done for STDIN:

$_ = <FILE>; or

while (<FILE>) {

statements;

}

To print to your open output file, use the filehandle as the first argument to the print statement (N.B. no commas between the filehandle and the string to print).

print FILE “Look Ma! No hands!\n”;

To change the default output filehandle from STDOUT to another one, use select:

select FILE;

From this point on, all calls to print or write without a filehandle argument will result in output being directed to the selected filehandle. Any special variables related to output will also then apply to this new default. To change the default back to STDOUT, select it:

select STDOUT;

When you are finished using a filehandle, close it using close():

close(FILE);

Formats:

Perl has a fairly sophisticated mechanism for generating formatted output. This involves using pictorial representations of the output, called templates, and the function write.

Using a format consists of three operations:

1. Defining the format (template)

2. Loading data to be printed into the variable portions of the format

3. Invoking the format.

Format templates have the following general form:

format FORMATNAME =

fieldline

$value_one, $value_two, …

Everything between the “=” and the “.” is considered part of the format and everything (in the fieldlines) will be printed exactly as it appears (whitespace counts). Fieldlines permit variable interpolation via fieldholders:

Hi, my name is @<<<<<, and I’m @< years old. <-- Fieldline

$name, $age <-- Valueline

Fieldholders generally begin with a @ and consist of characters indicated alignment/type.

@<<< Four character, left-justified field

@>>> Four character, right-justified field

@||| Four character, centered field

@###.## Six character numeric field, with two decimal places

@* Multi-line field (on line by itself – for blocks of text)

^<<<< Five character, “filled” field (“chews up” associated variables)

The name of the format corresponds to the name of a filehandle. If write is invoked on a filehandle, then the corresponding format is used. Naturally then, if you’re printing to standard output, then your format name should be STDOUT. If you want to use a format with a name other than that of your desired filehandle, set the $~ variable to the format name.

There are special formats which are printed at the top of the page. If the active format name is FNAME, then the “top” format name is FNAME_TOP. The special variable $% keeps a count of how many times the “top” format has been called and can be used to number pages.

Manipulating files & directories:

The action of opening a file for writing creates it. Perl also provides functions to

manipulate files without having to ask the operating system to do it.

unlink(filename)

Delete an existing file. Unlink can take a list of files, or wildcard as an argument as well: unlink(<*.bak>)

rename(oldname, newname)

This function renames a file. It is possible to move files into other directories by specifying a path as part of the new name. Directories also have some special function associated with them

mkdir(dirname, mode)

Create a new directory. The “mode” specifies the permissions (set this to 0777 to be safe).

rmdir(dirname)

Removes (empty) directories

chdir(dirname)

Change current working directory to dirname, File and directory attributes can be modified as well:

chmod(permission, list of files)

Change the permissions of files or directories:

666 = read and write

444 = read only

777 = read, write, and executable

utime(atime, mtime, list of files)

Modify timestamps on files or directories. “atime” is the time of the most recent access, and “mtime” is the time the file/directory was last modified.