Are some data structures more suitable for functional programming than others?

Question

In Real World Haskell, there is a section titled \"Life without arrays or hash tables\" where the authors suggest that list and trees are preferred in functional programming, w

Answer 1

Yes, the primary difference is immutability of the data, which can include code (see higher order functions). See the Wikipedia page on Purely Functional for a list of the common data types and usages. Whether its a problem or not depends on how you look at it. There are many advantages to programming in a functional language if it fits the type of task you are working on. A hash table is a type of associative array, but isn't the one you want to use in a functional language because of the rehashing you'd have to do on insert and the poor performance without arrays. Instead, try the Haskell implementation of Data.Map for an associative array.

Answer 2

The book Purely Functional Data Structures covers your questions in depth, and includes a great mix of theory and implementations primarily in ML - the appendix also contains Haskell implementations so you should be able to follow along with a bit of extra page turning. It is a pretty good (though difficult in parts) read if you are really interested in a thorough answer to your questions. Having said that I think ephemient gave a superb short answer.

edit: Steven Huwig provided a link to the thesis that the book started as. While I haven't read through it the only big thing missing (judging from the table of contents) are the Haskell implementations.

Answer 3

Chris Okasaki's thesis, Purely Functional Data Structures, is available for free online. It covers many different strategies for immutable persistent data representation.

As far as really needing a hash table, consider that an O(lg n) lookup is only twenty times as slow as an O(1) lookup when you are searching a million elements.