Click here to close now.


Java IoT Authors: Pat Romanski, Liz McMillan, Yeshim Deniz, SmartBear Blog, Elizabeth White

Related Topics: Java IoT

Java IoT: Article

Java Collections

Managing collections

As Jason Bell pointed out in his editorial "A Modern Day Cinderella" (JDJ, Vol. 8, issue 9), the spotlight is on J2EE and as a result many programmers are ignoring the foundation of the JDK. J2SE is the Java equivalent of C/C++ standard libraries. Here we deal with the lower-level entities, like the Number types, Integer, Long, Float, and Double.

The Java Collections Framework (JCF) should be your first choice when faced with the task of managing any type of collection. The Collections API is one of the most useful parts of the JDK. Looking back at the projects I've worked on over the past 13 years, to some degree all of them involved managing collections of data structures.

In this article I'll review the collections architecture. I'll also point out some of the useful features of the collections API (sorting and searching). To begin I'll go over the class categories, followed by a more detailed explanation of each. I encourage you to review Sun's documentation at

There are explicit class categories in the Collections Framework. The J2SE Collections Framework consists of interfaces, abstract base classes, and concrete implementations that provide a rich set of functionality for us. The implementations are the classes your application should be utilizing behind the scenes. There are implementations based on maps and others that are backed by arrays. You can make your collection read-only or you can add support for multithreaded access. How is a programmer to decide which entity to use? There are two main criteria: thread safety and usage semantics.

Usage semantics can be further broken down into collection- or map-based access. The library makes a distinction. The Map interface is not related to any of the Collection interfaces, because its main purpose in life is to map a key to an object, while the collection is just a loosely associated group of objects.

Figure 1 provides a class diagram showing the interfaces that make up the Collections API. The interfaces represent the ideal types you should be passing around in your application.


I strongly urge you to expose only the interfaces to clients of your classes. If you don't do this and instead pass around references to the concrete implementations, your code will become brittle due to the number of changes required to swap out one interface implementation for another. You should strive to expose the most general interface. For example, if a method is to return an ArrayList, first look and see if the methods exposed by the Collection interface will meet the needs of the intended usage (see Table 1). By doing this, you give yourself the opportunity to modify your method to return a LinkedList or any other type supporting the Collection interface. Who knows? You may even want to provide your own implementation of a Balanced Tree, and if you are instantiating and passing around references to a TreeMap, you'll have to alter the code at each reference.


The semantics of Sets are close to those of Lists. However, Sets lack the notion of direct random access. A Set is just a collection of objects that you may iterate over. A useful feature of Sets is that they do not allow duplicates as long as you override hashCode() and equals() from Object. Listings 1-3 provide a short program that will illustrate this effect. There is the main HashSetExample and two Person classes: one that does not override Object.equals()/hashCode() and one that does.

Running this program produces the follow output.

[000-11-1111, 222-23-1234, 000-11-1111]
[000-11-1111, 222-23-1234]

To remove duplicates your classes must override equals() from java.lang.Object. According to the Javadoc, overriding Object.hashCode() has more to do with performance. Interestingly, Sun's "Introduction to the Collections Framework Short Course" mentions overriding only the Object.hashCode(). Beware that if you follow the tutorial to the letter, you'll still have duplicate entries. You must override Object.equals(), as I've done in Listing 2, to prevent duplicates in your Sets.

How about sorting this list? TreeSet can do that for us, but we still have a choice to make. Will we be sorting by the natural order or do we want an ad hoc sort? For this article we'll examine the ad hoc sort (to implement your own natural order, your class should implement the Comparable interface). We can achieve an arbitrary sort order by utilizing the Comparator interface. When we employ a comparator, it's passed to the sorting object. First, we need to create our sorting algorithm (see Listing 4).

Now we can give this algorithm to the other implementation of Set; TreeSet. We add the following code to our main method at line 29 in Listing 4.

30 Set sortedSet =
31 new TreeSet
32 (new PersonComparator());
34 sortedSet.addAll(set);
36 // sorted
37 System.out.println(sortedSet);

Now the output becomes:

[000-11-1111, 222-23-1234, 000-11-1111]
[000-11-1111, 222-23-1234]
[222-23-1234, 000-11-1111]

Cool, eh? I won't go over this for each implementation. You should be able to apply this concept to any of the other sorting containers or utility methods (from Arrays or Collections). It's worth pointing out that even if you don't override either method, the TreeSet will use the Comparator and eliminate duplicates in the sorted set. Figure 2 provides a class diagram for the Set category.


The LinkedHashSet is a special implementation of HashSet that supports list operations without directly implementing the List interface. LinkedHashSet will maintain the insertion order of the list elements, yet still allow you to access elements via a key, such as a traditional Map. And, as the name implies, it is a Set that supports all of Set's operations.

Collection's other category is List; the implementations are ArrayList and LinkedList (see Figure 3).


The List interface supports the notion of direct index-based access to the entries, allows duplicates, and defines an order. Direct index-based access is realized via the get(int) method, which accepts an index as the only argument. You may even acquire a subset of the List by specifying a "from index" and a "to index", the semantics of which closely follow that of String. The element at "from index" will be included in the sublist while the element at "to index" will not.

ArrayList should be preferred if you don't require the ability to insert elements into the middle of the List (you're always adding to the end of the List) and you require random access to the elements. However, if you need to insert elements into the List and sequential access is your main concern, then LinkedList will be better.

Finally, we get to the Map category (see Figure 4). As mentioned earlier, Map is not related to any of the Collection classes. This is because the JCF authors wanted to make a clear distinction between Collections and Maps. The most notable difference is that Maps do not support index-based access semantics. What is the nth element of a Map?


If a Map is a Collection, what are the elements? The only reasonable answer is "Key-value pairs," but this provides a very limited (and not particularly useful) Map abstraction. You can't ask what value a given key maps to, nor can you delete the entry for a given key without knowing what value it maps to.

The workhorse of this category is HashMap. For inserting, deleting, and accessing elements, HashMap offers the best implementation. TreeMap is the sorted version and offers the ability to traverse the contents of the Map in a determined order.

As with the HashSet earlier, HashMap will require you to override Object.equals() and have a defined Object.hashCode() implementation on your own classes. And, of course, the objects you place in TreeMap should be comparable [or you must use the TreeMap(Comparator) constructor].

As with Sets, there's a special implementation of Map that supports a List-like view. LinkedHashMap provides for the same deterministic ordering as LinkedHashSet and supports all Map operations.

There's a another specialized implementation of Map, WeakHashMap, that uses weak references. By employing WeakReference, the garbage collector is able to destroy objects despite the Map's reference. If no other thread holds a reference to a key in the WeakHashMap, the garbage collector is free to collect the key-value pair.

The framework offers several opportunities for creating your own collection classes. The abstractions are for those instances where you want a more application-specific collection. There are several abstract classes implementing the interfaces with enough basic functionality to make your task less painful (see Figure 5).


In general, you won't be extending these classes unless you want to try some new algorithm or storage technique. Most likely you should turn your attention to the wrapper classes as implemented by the Collections class. Using the Decorator pattern, as these classes do, you may create highly specialized versions of the containers. There's an excellent example in the group of classes created by Piet Jonas for detecting type errors. Using Piet's classes, it's possible to have an exception thrown if an incorrect type is inserted into a collection. These classes employ the exact same design as the specialized wrappers available in the synchronization and read-only methods that I'll discuss next.

java.util.Collections API
Did you know that many of the Vector's methods are declared with the synchronized modifier? Are you aware of the cost of synchronization? While there have been advancements in many JVMs, there is still a slight overhead incurred with synchronization.

Unless several different threads might access your collection, forget about any of the thread-safe implementations. Use one of the nonthread-safe implementations, like ArrayList or HashMap. If you need index-based access, use the ArrayList. If you are more concerned about key-based access, use the HashMap.

While I may mention Vector and Hashtable from time to time, you should be aware that these two classes are now referred to as legacy code. The API has been reworked of late and all of the collection APIs are now unsynchronized. Special synchronized wrappers have been implemented (and hidden from us) for creating polymorphic, thread-safe implementations of the unsynchronized classes. You gain access to these thread-safe versions via static methods on the Collections class.

Collection Collections.synchronizedCollection(Collection);
List Collections.synchronizedList(List);
Map Collections.synchronizedMap(Map);
Set Collections.synchronizedSet(Set);
SortedMap Collections.synchronizedSortedMap(SortedMap);
SortedSet Collections.synchronizedSortedSet(SortedSet);

Notice that all of these methods accept the most general interface and return the same interface. If you make judicious use of these generalities, you'll be able to swap out implementations relatively painlessly. Now keep in mind that in theory, the implementation of collections shall be free to do whatever it wants. You don't want your code dependent upon J2SE source. If you insist on using the concrete classes, you'll have to downcast to use the results from the previous methods. Downcasting requires knowledge of implementation. Things will change over time. Try to insulate yourself from potential change points. The entire Collections Framework wreaks polymorphism, so take advantage of it, as polymorphism is a good thing.

I performed a small test to compare ArrayList, SynchronizedList, and Vector, all three of which implement the List interface. The results show that for synchronized updates, Vector is the worst performer, while SynchronizedList is much faster. Both are compared to the unsynchronized ArrayList. The test involved completing a read (get) or write (add) operation in a tight loop, 100,000 times (see Table 2).


Comparing Vector to SynchronizedList shows that Vector takes 138% and 40% more time than the same operations on SynchronizedList. Meanwhile, SynchronizedList takes a 12% hit over ArrayList for read operations, compared to the 167% increase for Vectors. Some people might be confused by the lack of symmetry in the numbers. If we want to compare A to B, the proper equation is (A - B)/B. Therefore if I want to compare 2 to 6, then (2 - 6)/6 gives -0.6667 or -66%. If I compare 6 to 2, then (6 - 2)/2 gives 2 or 200%. This may seem counterintuitive to saying 6 is three times as large as 2 (which is just a simple ratio, not a comparison).

All of the collections support iterator semantics. Some will bark at you if the underlying collection is altered while you are accessing the iterator by throwing a ConcurrentModification exception.

The static class Collections has many other useful methods for converting to and from certain types of collections. Of interest are those dealing with the creation of unmodifiable collections. First you create your collection and then pass it into the appropriate method and your collection is transformed into something that looks just like the original, but now it will throw an exception if anyone attempts to add or delete an object. Inner classes in Collections that simply extend the standard collection class and override the modifiers accomplish this. Now you can implement the Command pattern and employ the concept of read-only parameters and return structures. In a language that deals exclusively with object references, that's a nice-to-have feature.

List Collections.unmodifiableList(List);
Map Collections.unmodifiableMap(Map);
Set Collections.unmodifiableSet(Set);
SortedMap Collections.unmodifiableSortedMap(SortedMap);
SortedSet Collections.unmodifiableSortedSet(SortedSet);

Sorting has been taken care of with a "tuned" implementation of Merge Sort. There are routines for sorting primitives and objects. You can implement classes that have a natural order by extending Comparable. If inheritance is at a premium, use the Comparator interface. C++ programmers will feel right at home with this idiom from the STL. There are even collections that are themselves sorted. SortedTree allows you to add objects that will be sorted on the fly. The API is so flexible that you can implement the natural order.

Other utility methods in collections have to do with searching a List. The Collections class offers two binary searching methods.

Object Collections.binarySearch(List list, Object key);
Object Collections.binarySearch(List list, Object key, Comparator comparator);

These two methods, one of which employs the natural sort order of the list and the other the ad hoc, run in log(n) time where n is the number of elements in the list. However, this is true only if the list passed in implements the RandomAccess interface. Otherwise, if the list does not implement RandomAccess and is large, the search will execute an iterator-based binary search, which according to the Javadoc will "perform O(n) link traversals and O(log n) element comparisons."

Figure 6 shows the big picture with the preferred extension points highlighted. We've discussed the general categories: Collection (Set, List) and Map. We've played around a little and have seen that to take full advantage of some collections, we have to override Object.equals and Object.hashCode. Also, we went over some of the performance tradeoffs of a couple of implementations. I should mention that there are other Collection APIs available to Java programmers. There's the popular JGL and the JDSL. I haven't looked at the JGL but I have played around with the academic version of the JDSL. The JDSL gives you all those nifty data structures you talked about in your junior year algorithms class.


There are some new collections available in the latest JCF: LinkedHashSet, LinkedHashMap, and IdentityHashMap. In general they are highly specialized versions of the core JCF classes.

This article should prompt you to take another look at the Collections Framework and, if you are lucky, you'll see something that fits with your current development task. If you are really lucky, perhaps you'll see something else in the J2SE libraries that you never knew existed, collection related or not. Unfortunately, there doesn't seem to be any J2SE champion at Sun, so you'll have to make an effort to scan through the API's Javadoc every so often and perhaps even the source code as well (there are some novel snippets in there).

More Stories By David McReynolds

David McReynolds has been programming for over 12 years and is currently employed by Daugherty Business Solutions as a consultant. He has an MS in computer science from Southern Polytechnic State University.

Comments (6) View Comments

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.

Most Recent Comments
Abelardo 02/20/04 01:36:24 PM EST

I would like to nominate this article for the "best of the year"

David McReynolds 02/13/04 12:53:04 PM EST

Thanks Bill. I know I ran the code so perhaps I missed it during the cut-and-paste operation.

Bill Sommers 02/12/04 01:32:56 PM EST

This is a very nice article. I did find what appears to be an error in the source code. The Person class is missing a getSSN() method. I fixed the problem by writing this method:

public Object getSSN() {
return this.toString();

Troll Fiddler 02/10/04 05:13:27 AM EST

Superbly written with lovely clear diagrams. Good work. Not many authors explain things this well.

As Einstein said "if you can''t explain your work to a 10 year old, you''re a charlatan." This author is no charlatan.

Bee 02/05/04 03:07:11 PM EST

Ditto! what Shilpi said,This author is very well organized and excelent detail!

shilpi 02/05/04 03:00:42 PM EST

One of the most articulate and intelligent publishings I have read in a long time. I look forward to reading and hearing more from this author.

@ThingsExpo Stories
SYS-CON Events announced today that Luxoft Holding, Inc., a leading provider of software development services and innovative IT solutions, has been named “Bronze Sponsor” of SYS-CON's @ThingsExpo, which will take place on November 3–5, 2015, at the Santa Clara Convention Center in Santa Clara, CA. Luxoft’s software development services consist of core and mission-critical custom software development and support, product engineering and testing, and technology consulting.
“In the past year we've seen a lot of stabilization of WebRTC. You can now use it in production with a far greater degree of certainty. A lot of the real developments in the past year have been in things like the data channel, which will enable a whole new type of application," explained Peter Dunkley, Technical Director at Acision, in this interview at @ThingsExpo, held Nov 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA.
NHK, Japan Broadcasting, will feature the upcoming @ThingsExpo Silicon Valley in a special 'Internet of Things' and smart technology documentary that will be filmed on the expo floor between November 3 to 5, 2015, in Santa Clara. NHK is the sole public TV network in Japan equivalent to the BBC in the UK and the largest in Asia with many award-winning science and technology programs. Japanese TV is producing a documentary about IoT and Smart technology and will be covering @ThingsExpo Silicon Valley. The program, to be aired during the peak viewership season of the year, will have a major impac...
Developing software for the Internet of Things (IoT) comes with its own set of challenges. Security, privacy, and unified standards are a few key issues. In addition, each IoT product is comprised of at least three separate application components: the software embedded in the device, the backend big-data service, and the mobile application for the end user's controls. Each component is developed by a different team, using different technologies and practices, and deployed to a different stack/target - this makes the integration of these separate pipelines and the coordination of software upd...
"Matrix is an ambitious open standard and implementation that's set up to break down the fragmentation problems that exist in IP messaging and VoIP communication," explained John Woolf, Technical Evangelist at Matrix, in this interview at @ThingsExpo, held Nov 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA.
You have your devices and your data, but what about the rest of your Internet of Things story? Two popular classes of technologies that nicely handle the Big Data analytics for Internet of Things are Apache Hadoop and NoSQL. Hadoop is designed for parallelizing analytical work across many servers and is ideal for the massive data volumes you create with IoT devices. NoSQL databases such as Apache HBase are ideal for storing and retrieving IoT data as “time series data.”
There are so many tools and techniques for data analytics that even for a data scientist the choices, possible systems, and even the types of data can be daunting. In his session at @ThingsExpo, Chris Harrold, Global CTO for Big Data Solutions for EMC Corporation, will show how to perform a simple, but meaningful analysis of social sentiment data using freely available tools that take only minutes to download and install. Participants will get the download information, scripts, and complete end-to-end walkthrough of the analysis from start to finish. Participants will also be given the pract...
Clearly the way forward is to move to cloud be it bare metal, VMs or containers. One aspect of the current public clouds that is slowing this cloud migration is cloud lock-in. Every cloud vendor is trying to make it very difficult to move out once a customer has chosen their cloud. In his session at 17th Cloud Expo, Naveen Nimmu, CEO of Clouber, Inc., will advocate that making the inter-cloud migration as simple as changing airlines would help the entire industry to quickly adopt the cloud without worrying about any lock-in fears. In fact by having standard APIs for IaaS would help PaaS expl...
WebRTC is about the data channel as much as about video and audio conferencing. However, basically all commercial WebRTC applications have been built with a focus on audio and video. The handling of “data” has been limited to text chat and file download – all other data sharing seems to end with screensharing. What is holding back a more intensive use of peer-to-peer data? In her session at @ThingsExpo, Dr Silvia Pfeiffer, WebRTC Applications Team Lead at National ICT Australia, will look at different existing uses of peer-to-peer data sharing and how it can become useful in a live session to...
SYS-CON Events announced today that ProfitBricks, the provider of painless cloud infrastructure, will exhibit at SYS-CON's 17th International Cloud Expo®, which will take place on November 3–5, 2015, at the Santa Clara Convention Center in Santa Clara, CA. ProfitBricks is the IaaS provider that offers a painless cloud experience for all IT users, with no learning curve. ProfitBricks boasts flexible cloud servers and networking, an integrated Data Center Designer tool for visual control over the cloud and the best price/performance value available. ProfitBricks was named one of the coolest Clo...
Organizations already struggle with the simple collection of data resulting from the proliferation of IoT, lacking the right infrastructure to manage it. They can't only rely on the cloud to collect and utilize this data because many applications still require dedicated infrastructure for security, redundancy, performance, etc. In his session at 17th Cloud Expo, Emil Sayegh, CEO of Codero Hosting, will discuss how in order to resolve the inherent issues, companies need to combine dedicated and cloud solutions through hybrid hosting – a sustainable solution for the data required to manage I...
Mobile messaging has been a popular communication channel for more than 20 years. Finnish engineer Matti Makkonen invented the idea for SMS (Short Message Service) in 1984, making his vision a reality on December 3, 1992 by sending the first message ("Happy Christmas") from a PC to a cell phone. Since then, the technology has evolved immensely, from both a technology standpoint, and in our everyday uses for it. Originally used for person-to-person (P2P) communication, i.e., Sally sends a text message to Betty – mobile messaging now offers tremendous value to businesses for customer and empl...
Scott Guthrie's keynote presentation "Journey to the intelligent cloud" is a must view video. This is from AzureCon 2015, September 29, 2015 I have reproduced some screen shots in case you are unable to view this long video for one reason or another. One of the highlights is 3 datacenters coming on line in India.
Nowadays, a large number of sensors and devices are connected to the network. Leading-edge IoT technologies integrate various types of sensor data to create a new value for several business decision scenarios. The transparent cloud is a model of a new IoT emergence service platform. Many service providers store and access various types of sensor data in order to create and find out new business values by integrating such data.
SYS-CON Events announced today that IBM Cloud Data Services has been named “Bronze Sponsor” of SYS-CON's 17th Cloud Expo, which will take place on November 3–5, 2015, at the Santa Clara Convention Center in Santa Clara, CA. IBM Cloud Data Services offers a portfolio of integrated, best-of-breed cloud data services for developers focused on mobile computing and analytics use cases.
Apps and devices shouldn't stop working when there's limited or no network connectivity. Learn how to bring data stored in a cloud database to the edge of the network (and back again) whenever an Internet connection is available. In his session at 17th Cloud Expo, Bradley Holt, Developer Advocate at IBM Cloud Data Services, will demonstrate techniques for replicating cloud databases with devices in order to build offline-first mobile or Internet of Things (IoT) apps that can provide a better, faster user experience, both offline and online. The focus of this talk will be on IBM Cloudant, Apa...
The enterprise is being consumerized, and the consumer is being enterprised. Moore's Law does not matter anymore, the future belongs to business virtualization powered by invisible service architecture, powered by hyperscale and hyperconvergence, and facilitated by vertical streaming and horizontal scaling and consolidation. Both buyers and sellers want instant results, and from paperwork to paperless to mindless is the ultimate goal for any seamless transaction. The sweetest sweet spot in innovation is automation. The most painful pain point for any business is the mismatch between supplies a...
As a company adopts a DevOps approach to software development, what are key things that both the Dev and Ops side of the business must keep in mind to ensure effective continuous delivery? In his session at DevOps Summit, Mark Hydar, Head of DevOps, Ericsson TV Platforms, will share best practices and provide helpful tips for Ops teams to adopt an open line of communication with the development side of the house to ensure success between the two sides.
As more and more data is generated from a variety of connected devices, the need to get insights from this data and predict future behavior and trends is increasingly essential for businesses. Real-time stream processing is needed in a variety of different industries such as Manufacturing, Oil and Gas, Automobile, Finance, Online Retail, Smart Grids, and Healthcare. Azure Stream Analytics is a fully managed distributed stream computation service that provides low latency, scalable processing of streaming data in the cloud with an enterprise grade SLA. It features built-in integration with Azur...
WebRTC: together these advances have created a perfect storm of technologies that are disrupting and transforming classic communications models and ecosystems. In his session at WebRTC Summit, Cary Bran, VP of Innovation and New Ventures at Plantronics and PLT Labs, will provide an overview of this technological shift, including associated business and consumer communications impacts, and opportunities it may enable, complement or entirely transform.