As mentioned above, table indexing is one-based by default. This means tables are created as follows:

 t = { 3,2,5,1,4 }

Tables returned by HyperMath functions are one-based indexed by default (unless noted otherwise). Tables are versatile enough to accommodate non-positive integers as indices. For example, the following are valid:

> t = {}

> t[0] = 3

> t[-1] = 2

The above can be useful in some situations, like using the data as indices to match corresponding values to avoid lookup tables. However, using non-positive integer indices should be avoided when possible.

In fact, indices for tables are more than just numeric entities; they are like keys in a database. Strings can also be used for indexing:

> t = {}

> t[0] = 3

> t["ID"] = 2

Hence, table indexing has no bounds checking, and invalid indexing results in a nil return.

Indices and corresponding entries can be accessed by the pairs() function as in:

> for i,v in pairs(t) do print(i,v) end

ID    2

0    3

There are two utility functions, GetTableIndices and GetTableKeys, that allow querying table indices by type.  See the topic Utility Functions > Operations.