AlphabeticIndex supports the creation of a UI index appropriate for a given language. It can support either direct use, or use with a client that doesn't support localized collation. The following is an example of what an index might look like in a UI:

 ... A B C D E F G H I J K L M N O P Q R S T U V W X Y Z  ...

 A
    Addison
    Albertson
    Azensky
 B
    Baecker
 ...

The class can generate a list of labels for use as a UI "index", that is, a list of clickable characters (or character sequences) that allow the user to see a segment (bucket) of a larger "target" list. That is, each label corresponds to a bucket in the target list, where everything in the bucket is greater than or equal to the character (according to the locale's collation). Strings can be added to the index; they will be in sorted order in the right bucket.

The class also supports having buckets for strings before the first (underflow), after the last (overflow), and between scripts (inflow). For example, if the index is constructed with labels for Russian and English, Greek characters would fall into an inflow bucket between the other two scripts.

Direct Use

The following shows an example of building an index directly. The "show..." methods below are just to illustrate usage.

// Create a simple index where the values for the strings are Integers, and add the strings

AlphabeticIndex<Integer> index = new AlphabeticIndex<Integer>(desiredLocale).addLabels(additionalLocale);
int counter = 0;
for (String item : test) {
    index.addRecord(item, counter++);
}
...
// Show index at top. We could skip or gray out empty buckets

for (AlphabeticIndex.Bucket<Integer> bucket : index) {
    if (showAll || bucket.size() != 0) {
        showLabelAtTop(UI, bucket.getLabel());
    }
}
 ...
// Show the buckets with their contents, skipping empty buckets

for (AlphabeticIndex.Bucket<Integer> bucket : index) {
    if (bucket.size() != 0) {
        showLabelInList(UI, bucket.getLabel());
        for (AlphabeticIndex.Record<Integer> item : bucket) {
            showIndexedItem(UI, item.getName(), item.getData());
        }

The caller can build different UIs using this class. For example, an index character could be omitted or grayed-out if its bucket is empty. Small buckets could also be combined based on size, such as:

... A-F G-N O-Z ...

Client Support

Callers can also use the AlphabeticIndex.ImmutableIndex, or the AlphabeticIndex itself, to support sorting on a client that doesn't support AlphabeticIndex functionality.

The ImmutableIndex is both immutable and thread-safe. The corresponding AlphabeticIndex methods are not thread-safe because they "lazily" build the index buckets.

• ImmutableIndex.getBucket(index) provides random access to all buckets and their labels and label types.
• AlphabeticIndex.getBucketLabels() or the bucket iterator on either class can be used to get a list of the labels, such as "...", "A", "B",..., and send that list to the client.
• When the client has a new name, it sends that name to the server. The server needs to call the following methods, and communicate the bucketIndex and collationKey back to the client.

  int bucketIndex = index.getBucketIndex(name);
  String label = immutableIndex.getBucket(bucketIndex).getLabel();  // optional
  RawCollationKey collationKey = collator.getRawCollationKey(name, null);
  

• The client would put the name (and associated information) into its bucket for bucketIndex. The collationKey is a sequence of bytes that can be compared with a binary compare, and produce the right localized result.