hash tables

Calling sequence

Parameters

• size: an integer, giving initial expected size of the hash table.
• var1,...,varn: any nsp objects having a name.
• val1: any nsp expression which evaluates to a nsp object.

Description
The hash data type is useful to maintain a collection of objects which can be accessed using their names as access key. A hash table could be created with the hash_create function (you can use either the shortly name +hash+). Internally nsp hash tables are implemented as hash tables with open adressing. They are dynamically resized according to the number of stored objects.

Operations on hashs

• .__keys return a string matrix containing all the keys present in the hash table.

• .delete[S1,...,Sn]: delete the entries whose keys are present in the string matrices S1,...,Sn.
• .enter[var1,...,varn,name1=val1,...]: enter new objects in the hash table with keys given by their names or by explicit naming (name=val). Note that if some keys already preexist in the hash table, their value are replaced by new ones.
• .find[S1,...,Sn]: returns on the calling stack for each key given by the values of the string matrices S1,...,Sn the corresponding stored objet if present or an error if one key is not present in the hash table. Thus, this method returns a variable number of values.
• .iskey[S1,...,Sn]: returns on the calling stack for each given entry key a boolean which states if the given key is present or not in the hash table.
• .merge[H]: insert the entries of hash table H in the hash table object.

• If ˆverb+key+ is a string H(key) returns the element of H given by key. This is therefore equivalent to H.find[key]. It is also possible to use the syntax H.name (Note that in that case name given the key name not an expression which evaluates to a string)
• H(:) puts all the elements of H on the calling stack but as named optional arguments. Thus is can be usefull to pass a set of named options to a function. (H=hash_create(color=3,size=5) and f(70,H(:)) is equivalent to f(70,color=3,size=5)).
• H(key)=e or H.key=e insert the rhs value in the hash table with key name given by key.

• length(H) return the number of objets stored in H
• size(H) return [m,n] where m stands for the number of objets stored in H and n is the size of internal size of the hash table.

Examples

A = rand(4,4); 
H = hash_create(5,A,B=34,C="foo")  
// or shortly H = hash(5,A,B=34,C="foo") 
H.delete['A',['B','C']]; 
H.enter[A,Z='foo',C=3.56]; 
[a,b,c]=H.find['A',['B','C']] 
[a,b,c]=H.iskey['A',['B','C']] 
H.merge[hash_create(P=89,Q=67)]; 
H.A = 56  
str = 'A' 
H(str) = 45; 
function y=f(a=1,b=2,c=3);y=[a,b,c];endfunction  
H = hash_create(a=67,b=45); 
f(H(:)); 

Authors jpc