# Difference between revisions of "INF, NAN, and Null"

These are special values that Analytica returns in particular conditions. You can also use them in expressions:

Inf means infinity -- e.g.,

`1/0 → Inf`

or a number larger than 1.796E308 (the largest number that your computer can represent explicitly) -- e.g.

`1E307 * 100 → Inf`

-Inf means negative infinity (or a number less than 1.796E308) -- e.g.

`-1/0 → -Inf`

NAN means "Not A Number" -- i.e. not a well-defined number nor infinity -- e.g.

`0/0 → NAN`
`Sqrt(-1) → NAN`

(If you enable Complex Numbers, `Sqrt(-1)` returns the valid imaginary number, 1j.)

Null means that there is no such value. For example, Slice and Subscript return `Null` if you try to get the nth slice over an Index with less than n values. For example:

`Index Year := [2015, 2016, 2017]`
`Slice(Year, 4) → NULL`
`Variable X := Array(Year, [20, 23, 28])`
`X[Year = 2018] → NULL`

### More on INF and NAN

Calculations using `INF` and `NAN` follow ANSI (Association of National Standards Institutes) standards, which follow the laws of mathematics as far as possible:

`1/Inf → 0`
`1/(-Inf) → 0`
`Inf + Inf → Inf`
`Inf - Inf → NAN`

Expressions taking `NAN` as an operand or parameter give `NAN` as their result unless the expression has a well-defined logical or numerical value for any value of `NAN`:

`True OR NAN → True`
`NaN AND False → False`
`IF True THEN 5 ELSE NAN → 5`

### More on NULL

When `Null` appears in scalar operations, it generally produces a warning and evaluates to `Null`, for example:

`10 + NULL → NULL`
`NULL - 10 → NULL`
`1 AND NULL → NULL`

Array-reducing functions ignore `Null`. These examples demonstrate (assume `A` is indexed by `I` as indicated).

`Variable A :=`
I ▶
1 2 3 4 5
8 NULL 4 NULL 0
`Sum(A, I) → 12`
`Average(A, I) → 4`
`JoinText(A, I, ', ') → "8, 4, 0"`

Graphs will simply ignore (not show) any point whose value is `Null`.

Array-reducing functions include Sum, Min, Max, ArgMin, ArgMax, Product, Average, JoinText, Irr, Npv. Array functions Sum, Min and Max also accept an optional parameter «IgnoreNaN» to ignore `NaN` values (which otherwise propagate, i.e. return `NaN`).

Regression also ignores any data points which have `Y = Null`, which is useful for missing data.