Softpanorama (slightly skeptical) Open Source Software Educational Society

May the source be with you, but remember the KISS principle ;-)

Introduction to Perl for Unix System Administrators

(Perl without OO and excessive complexity)

by Dr Nikolai Bezroukov

Chapter 3: Perl Data Structures

Version 0.72(beta)

Contents

3.1. String operations

3.2. Operations on Arrays

3.3. Operations on Associative Arrays

3.4. Summary

3.5. Homework

Perl supports several types of data structures: 

• strings (they are usually called scalars)
• arrays (actually this not an array in C language style -- it is a list that behaves much like text editor buffer.  You can get lines from the buffer using line numbers, delete lines, exchanges lines, etc. The first line has index (line number) 0(zero); there is no principal differences between arrays and associative array (hashes, see below) in Perl -- both are indexed lists, but arrays are more close to text editor buffer paradigm, while hashes are closer to symbol table paradigm)
• associative arrays or hashes. An associative is like an array in that it stores strings in sequence but uses another string as an index to address individual items instead of a number. In essence this an abstraction of compiler symbol table. Implementation-wise binary search or hashing functions are involved.
• files and pipes (we will discuss them later in a separate chapter)

Another way to look into data structures is to analyze their strength and weaknesses

• String -- fast access to substring, slower access to a random element. Rather slow search of substring especially if length is big (in kilobytes or megabytes). Good access to sequential elements in both direction. Any change involves freeing and reallocating memory. 
• Array (list or, more correctly, text editor buffer)-- fast access to element if index is known, reasonably quick insertion, reasonable deletion. Built-in sort operation exist, so ordered array provides the possibility to speed search using binary search or other algorithms. 
• Hash table (associative array) -- fast search (if the key is known), fast insertion. Fast deletion (key should be known). With fast insertion hash arrays are somewhat similar to lists and they can be converted to lists. Typically hashes are converted to lists in foreach loops.

All in all Perl provide a rich set of built-in operation for lists and associative arrays (hashes).  It also provides some connections between these two data structures. Here Perl is really one of the top if not the top language in existence. If you task can be structured of information along these lines (text processing) you can immensely benefit from using Perl.

At the same time multidimensional arrays are not natural to Perl and if your task does require them you probably will be in disadvantage and should consider either using references or switch to other language that supports this abstraction better.

Perl warts

While in many way a very innovative language Perl suffers from some design decisions connected with using separate sets of build-in functions for strings and arrays:

The main problem with Perl that I see is the although it provides a rich set of functions for string and arrays they are not generalized in such a way that similar functions are applicable to both without renaming, although this is entirely possible. So instead of substr functions (that is defined on strings) you need to use splice function for arrays and remember idiosyncrasies of two functions instead of one. Here Larry Wall made a very bad decision, but we can do nothing about it. You will see the problem yourself after you will face the assortment of functions that are available in Perl for arrays and strings. Actually you are forced to use a subset as full assortment is beyond human capacities to master (starving in the food store effect ;-).

Created: November 7 1998; Last modified: March 15, 2008