Stream (abstract data type)

For the more applied concept in computing, see Stream (computing).

In type theory and functional programming, a stream is a potentially infinite analog of a list, given by the coinductive definition:^[1][2]

data Stream α = Nil | Cons α (Stream α)

Generating and computing with streams requires lazy evaluation, either implicitly in a lazily evaluated language or by creating and forcing thunks in an eager language. In total languages they must be defined as codata and can be iterated over using (guarded) corecursion.

UML package diagram of the stream hierarchy in .NET

Java provides the Stream interface under the java.util.stream namespace.

Python have the StreamReader and StreamWriter classes in the asyncio module.^[3]

.NET provides the abstract class Stream^[4] which is implemented by classes such as FileStream and MemoryStream.^[5]

See also

• Coinduction

References

1. Yamagiwa, Shinichi (2022), Katoh, Naoki; Higashikawa, Yuya; Ito, Hiro; Nagao, Atsuki (eds.), "Stream-Based Lossless Data Compression", Sublinear Computation Paradigm: Algorithmic Revolution in the Big Data Era, Singapore: Springer, pp. 391–410, doi:10.1007/978-981-16-4095-7_16, ISBN 978-981-16-4095-7
2. Sangma, Jerry W.; Sarkar, Mekhla; Pal, Vipin; Agrawal, Amit; Yogita (2022-04-01). "Hierarchical clustering for multiple nominal data streams with evolving behaviour". Complex & Intelligent Systems. 8 (2): 1737–1761. doi:10.1007/s40747-021-00634-0. ISSN 2198-6053.
3. "Streams". Python documentation. Retrieved 17 December 2024.
4. "Stream Class (System.IO)". learn.microsoft.com. Microsoft. Retrieved 17 December 2024.
5. "File and Stream I/O - .NET". learn.microsoft.com. Microsoft. 15 September 2021. Retrieved 17 December 2024.