|
D.2 Cluster Arrays (collections) |
SheerPower Clusters
Cluster Arrays (Collections)
A cluster array in SheerPower is similar to an in-memory spreadsheet, containing rows and columns of data. These clusters can efficiently store and manipulate data, making them ideal for handling information like spreadsheets.
SheerPower includes a rich set of features that allows you to input spreadsheet files directly into clusters and perform various operations such as sorting, including, excluding, and searching much like you would in a database.
Main Features of SheerPower Cluster Arrays
- Dynamic Key-Value Manipulation: SheerPower clusters support dynamic key-value manipulation natively. All fields in a cluster are mutable (changeable), which allows for flexible data handling.
- Multiple Key Fields: Clusters can have multiple key fields, which can be dynamically created as needed.
- Duplicate Values: Clusters allow for duplicate values in their fields, providing flexibility in data storage.
- Dynamic Resizing: Cluster arrays in SheerPower are dynamically sized, meaning they automatically adjust their size as rows are added or removed. This dynamic resizing is handled by SheerPower, ensuring efficient memory management without requiring manual intervention by the developer.
- Performance-Oriented: SheerPower clusters are optimized for performance, particularly in environments with abundant RAM. They utilize internal caching and hinting to reduce overhead, ensuring high efficiency.
Creating a Cluster Array to Hold Student Information
Let's create a cluster array called STUDENT to hold information about
students. Each row in this cluster will represent a different student, and we will track
each student's name, age, and grade level.
The most common method to add data to a cluster is by adding a new row to the end of the cluster using the ADD CLUSTER statement.
In this case, a new cluster row is established, storing information about the student "Joan Ark," who is 18 years old and in grade level 12.
Reset cluster and delete from cluster for details.
To remove all rows from a cluster — commonly used when reloading data — use:
delete from cluster xxx: all
where xxx is the cluster name.
Adding More Students
The row added last is considered the current row. For example, after the above code, row three (Desmond Jones) is current.
To print information about row one, you first make row one current:
Working with Cluster Arrays
To find out how many rows are in a cluster, use the SIZE() function:
To ask which row is current, use the ASK CLUSTER statement:
Cluster-Related Functions
SheerPower provides built-in functions for efficiently handling cluster arrays. These functions retrieve key information about a cluster’s data without requiring manual iteration.
| Function | Description |
|---|---|
size(cluster) |
Returns the total number of rows in the cluster. |
min(cluster->var) |
Returns the smallest value in the specified column. |
max(cluster->var) |
Returns the largest value in the specified column. |
sum(cluster->var) |
Returns the sum of all values in the specified column. |
Example Usage:
These functions optimize performance by scanning all cluster rows and extracting only the values for the specified variable. This eliminates the need for manual loops, making data operations very efficient.
Iterating Through Cluster Arrays
To operate on each row in a cluster array, use the COLLECT and END COLLECT statements.
For example, to print the name, age, and grade level of each student and
calculate their average age:
Sorting and Filtering Data
SheerPower makes it easy to filter and sort data within a cluster array. Here's how you can do it:
By default, the collect cluster statement processes the
cluster from the first row to the last. If you need to reverse this
order, use the reverse option, which reads from the last
row to the first. The syntax for this option is:
Steps for Using collect cluster
- Begin by collecting student records using the
collectandend collectblock. - Within this block, filter data using
includeandexcludestatements. For example, you can include only students over 16 years old. - Place any
sortstatements after the filtering step. The default sort order is ascending. - To sort in descending order, use the
sort descendingstatement. - You can include multiple
include,exclude, andsortstatements within the collect block to refine your data further.
Once the data is collected and sorted, you can iterate through the records using the
for each statement:
To include only students older than 16:
Searching Cluster Arrays
Use the FINDROW() function to search a cluster array for specific data:
The FINDROW() function is highly optimized, capable of
performing over 10 million searches per second, making it ideal for tasks requiring fast lookups.
In SheerPower, when you use the findrow() function with
specific fields in a cluster, those fields are automatically optimized
and treated as key fields. These key fields are managed
dynamically, meaning:
- No Unique Value Requirement: The values in these key fields do not need to be unique. Multiple rows can have the same key field value.
- Editable Fields: You can modify the content of these key fields at any time without restrictions, providing flexibility in organizing and accessing your data.
- Automatic Optimization: SheerPower automatically handles the optimization of these fields, so developers do not need to define or manage key fields manually.
- Flexible Data Types: Any data type, whether numeric or string, can be used as a key field.
This all enables efficient searching and manipulation of data in clusters without requiring developers to predefine or rigidly maintain key fields.
Summary: Cluster Arrays in SheerPower
Cluster arrays in SheerPower act like in-memory spreadsheets, organizing data into rows and columns. They simplify data management with features like dynamic resizing, key-value manipulation, and efficient searching. Cluster arrays can be extremely large. If your computer has sufficient RAM and pagefile space, clusters can support up to one billion rows with key lookup speeds of millions per second.
Key Features:
- Dynamic Resizing: Automatically adjusts as rows are added or removed.
- Fast Searching: Supports millions of searches per second
with
findrow(). - Sorting and Filtering: Built-in tools for refining data with include, exclude, and sort options.
Cluster arrays combine flexibility and performance, making them essential for efficient data handling in SheerPower applications.
