Input and Output

We'll finish up today with two topics that initially might not seem to fit with everything else we've talked about concerning scalar data: handling simple input and output. I've included them here essentially for one reason: so you know what's been going on in the scripts you've been writing that read input from the keyboard and print output to the screen.

In this section we'll talk about simple input and output, and as the book progresses you'll learn more about input and output to disk files, culminating on Day 15, “Working with Files and I/O.”

File Handles and Standard Input and Output

First, some terminology. In the scripts you've been looking at today and yesterday, you've used Perl code to read input from the keyboard and to write output to the screen. In reality, the keyboard and the screen aren't the best terms to use because, actually, you're reading from a source called standard input, and writing to a destination called standard output. Both of these concepts are borrowed from Unix systems, where using pipes and filters and redirection are common, but if you're used to Windows or the Mac, the idea of a standard input or output might not make much sense.

In all cases, when you're reading data from a source, or writing data to a destination, you'll be working with what are called file handles. Most often, file handles refer to actual files on the disk, but there are instances where data might be coming from or going to an unnamed source, for example, from or to another program such as a Web server. To generalize data sources and destinations that are not actual files, Perl gives you built-in file handles for standard input and standard output called STDIN and STDOUT (there's also STDERR, for standard error, but we'll leave that for later). These two file handles happen to include (and, in fact, are most commonly used for) input from the keyboard and output from the screen.

Reading a Line from Standard Input with <STDIN>

In the scripts we've seen so far in this book, there's usually been a line for reading input from the keyboard that looks something like this:

chomp($inputline = <STDIN>);

You'll see that line a lot in Perl code, although often it occurs on multiple lines, something like this (the two forms are equivalent):

$inputline = <STDIN>;
chomp($inputline);

You know now that $inputline is a scalar variable, and that you're assigning something to it. But what?

The STDIN part of this line is the special built-in file handle for standard input. You don't have to do anything to open or manage this special file handle; it's there for you to use. In case you're wondering why it's in all caps; that's a Perl convention to keep from confusing file handles from other things in Perl (such as actual keywords in the language).

The angle brackets around <STDIN> are used to actually read input from a file handle. The <> characters, in fact, are often called the input operator. <STDIN>, therefore, means read input from the STDIN file handle. In this particular case, where you're assigning the <STDIN> expression to a scalar variable, Perl will read a line from standard input and stop when it gets to a newline character (or a carriage return on the Macintosh). Unlike in C, you don't have to loop through the output and watch every character to make sure it's a newline; Perl will keep track of that for you. All you need is <STDIN> and a scalar variable to store the input line in.

All this talk about input and output brings us to the somewhat amusingly named chomp function. When you read a line of input using <STDIN> and store it in a variable, you get all the input that was typed and the newline character at the end as well. Usually, you don't want that newline character there, unless you're printing the input right back out again and it's useful for formatting. The built-in Perl chomp function, then, takes a string as input, and if the last character is a newline, it removes that newline. Note that chomp modifies the original string in place (unlike string concatenation and other string-related functions, which create entire new strings and leave the old strings alone). That's why you can call chomp by itself on its own line without reassigning the variable that holds that string.

Writing to Standard Output with print

When you get input into your Perl script with <STDIN>, or from a file, or from wherever, you can use Perl statements to do just about anything you like with that input. The time comes, then, when you'll want to output some kind of data as well. You've already seen the two most common ways to do that: print and printf.

Let's start with print. The print function can take any number of arguments and prints them to the standard output (usually the screen). Up to this point we've only used one argument, but you can also give it multiple arguments, separated by commas. Multiple arguments to print, by default, will get concatenated together before they get printed:

print 'take THAT!';
print 1, 2, 3;    # prints '123'
$a = 4;
print 1, ' ', $a;  # prints "1 4"
print 1, " $a";    # same thing

I mentioned the STDOUT file handle earlier, as the way to access the standard output. You might have noticed, however, that we've been printing data to the screen all along with print, and we've never had to refer to STDOUT. That's because Perl, to save you time and keystrokes, assumes that if you use print without an explicit file handle, you want to use standard output. In reality, the following Perl statements do exactly the same thing:

print "Hello World!
" ;

print STDOUT "Hello World!
";

More about the longer version of print when you learn more about file handles that are attached to actual files, on Day 15.

printf and sprintf

In addition to the plain old workhorse print, Perl also provides the printf and sprintf functions, which are most useful in Perl for formatting and printing numbers in specific ways. They work almost identically to those same functions in C, but beware: printf is much less efficient than print, so don't just assume you can use printf everywhere because you're used to it. Only use printf when you have a specific reason to do so.

As you learned yesterday, you use the printf function to print formatted numbers and strings to an output stream, such as standard output. sprintf formats a string and then just returns that new string, so it's more useful for nesting inside other expressions (in fact, printf calls sprintf to do the actual formatting).

Both printf and sprintf take two or more arguments: the first, a string containing formatting codes, and then one or more values to plug into those codes. For example, we've seen examples of printf that rounded off a floating-point number to two decimal places, like this:

printf("Average (mean): %.2f", $avg);

We've seen one that truncates it to an integer, like this:

printf("%d degrees Celsius
", $cel);

Yesterday, you also saw how to use sprintf to round a floating-point number to two digits of precision:

$value = sprintf("%.2f", $value);

The format codes follow the same rules as the C versions (although the * length specifier isn't supported), and can get quite complex. A simple formatting code that you might use in Perl looks like this:

%l.px

The x part is a code referring to the type of value; in Perl you'll be most interested in the d formatting code for printing integers, and the f formatting code for printing floating-point numbers. The l and the p in the formatting code are both optional. l refers to the number of characters the value should take up in the final string (padded by spaces if the value as printed is less than l), and p is the number of digit precision of a floating-point number. All numbers are rounded to the appropriate precision.

If you need to print an actual percent sign in your output, you'll need to use two of them:

printf("%d%% humidity 
", $hum);

Here are some typical examples of how either sprintf or printf might be used:

$val = 5.4349434;
printf("->%5d
", $val);     # 5
printf("->%11.5f
", $val);  # 5.43494
printf("%d
", $val);        # 5
printf("%.3f
", $val);      # 5.435
printf("%.1f
", $val);      # 5.4

Multiple formatting codes are interpolated left to right in the string, each formatting code replaced by an argument (there should be an equal number of formatting codes and extra arguments):

printf("Start value : %.2f End Value: %.2f
", $start, $end);

In this example, if $start is 1.343 and $end is 5.33333, the statement will print this:

Start value : 1.34 End Value: 5.33

If you're unfamiliar with C's printf formatting codes, you might want to refer to the perlfunc man page (or the printf man page) for more details.