Type and Shape

Even though ELI does not require explicit type for declaring a variable, it needs types for every computation internally. Moreover, a variable with a specific type at one time is assigned with a shape as well.

ELI Types

Type Value Symbol Null* Example
Boolean 0 B 0n 0,1
Integer 1 I 0n 123
Float 2 E 0nf 2.5
Char 3 C ‘abc’
Complex 4 X 0nx 1.5j2
Symbol 5 S ` `ibm
List 6 L () (`chen`wu;10 20)
Month 7 M 0nm 2012.08m
Date 8 D 0nd 2012.08.15
Datetime 9 Z 0nz 2012.08.15T13:25:17.357
Minute 10 U 0nu 13:25
Second 11 V 0nv 13:25:17
Time 12 T 0nt 13:25:17.357
Enumeration 13 N 0ne ...
Dictionary 14 Y 0ny ...
Keyed Table 15 K 0nk ...
Table 16 A 0na ...

* Note: Currently, null value is only available in Boolean, Integer, Float, Char, Symbol and List.

The types can be classified into the following two groups.

  1. Basic types:

    boolean, integer, float, char, complex, symbol, month, date,datetime, minute second, time

  2. Advanced types:

    list, enumeration, dictionary, keyed table and table

ELI follows some conventional type rules while new type rules because of new advanced types. ELI does not implicitly do casting for the result after one computation. Instead, ELI leaves it to developers that makes the system more consistent.

  • The arithmetic operations between boolean, int and double are the same as common languages such as C and Java.
  • Primitive functions like circle can generate complex numbers even input data is not a complex number. For example, _1*.0.5 (equivalent to sqrt(-1)) is 0j1 in ELI.
  • Table and dictionary can be transformed from each other if conditions are satisfied. (Check the primitive function transpose &.)

ELI Shapes

There are 7 basic shapes for basic types.

Shape Example Description
scalar x <- 2 shape(x) => empty
one-item vector x <- ,2 shape(x) => 1 (1-dim)
vector x <- 1 2 3 shape(x) => 3 (1-dim)
matrix x <- 2 2#1 2 3 4 shape(x) => 2 2 (2-dim)
high-dimension x <- 2 2 2#!8 shape(x) => 2 2 2 (3-dim)
list x <- (1 2; 2.5 9; ‘hello’) shape(x) => 3 (the length of the list)
table x <- ([]int<-1 2;flt<-2.3 5) shape(x) => 2 (row of the table)

* Note: For advanced types, they are formed by list and table shapes. And the two shapes are based on other basic types.