Partition methods for collections in C#

The article explains partitioning as splitting a collection into chunks of a fixed size, with the last chunk possibly smaller than the others. It gives an example of a 100-item list split into partitions of 30 items, producing four parts.

Reading notes

#

• Partitioning is defined as splitting a collection into partitions of size C, where C is a positive constant integer.
• A list of 100 elements partitioned by 30 yields four partitions, and the last partition contains only 9 elements.
• The article shows extension methods named Partition for IList, IDictionary<TKey, TValue>, and ConcurrentDictionary<TKey, TValue>.
• Each method uses Skip and Take inside a loop based on the collection count divided by the partition size.
• The example partitions a ConcurrentDictionary with 200,000 key-value pairs into four parts of 50,000 items each.
• The resulting partitions are processed in parallel.
• The article notes that the simpler Parallel.ForEach approach tested faster than partitioning into multiple concurrent dictionaries.
• The main point is that partitioning can support a divide-and-conquer approach for distributing collection work across several threads.