# LOOKUP
URL: https://docs.valsight.ai/lookup/
Description: Performs an approximate match lookup in one column and returns the corresponding value from another column.
**Category:** [Filtering & data shaping](/filtering-data-shaping/)

## Overview

The **LOOKUP** function **performs an approximate match lookup in one column and returns the corresponding value from another column**.

Use this function when you need threshold-based or approximate mapping from one value range to another.

## Syntax

`LOOKUP('Node1', 'Node2', 'Node3' [, DefaultReturnValue] [, "LevelName"])`

## Parameters

| Parameter              | Description                                                                                                                | Type           | Required | Default      |
| ---------------------- | -------------------------------------------------------------------------------------------------------------------------- | -------------- | -------- | ------------ |
| **Node1**              | **Lookup Value**. The value to search for in the index table.                                                              | Node reference | Yes      | --           |
| **Node2**              | **Index**. A column of data that is sorted in ascending order. The `LOOKUP` function searches for the value in this range. | Node reference | Yes      | --           |
| **Node3**              | **Return Value**. A range that contains the result.                                                                        | Node reference | Yes      | --           |
| **DefaultReturnValue** | The value returned when a lookup value is not found.                                                                       | Number         | No       | `-1`         |
| **LevelName**          | The level along which the index values are matched.                                                                        | Level name     | No       | `'Sequence'` |

## Watch out

* **Important:** the values in the index must be placed in **ascending order (alphabetical sorting)**, otherwise `LOOKUP` might not return the correct value.
* The `Index` values have to be sorted as `001`, `002`, `003`, ... instead of `1`, `2`, `3`, ... Otherwise, Valsight will consider, for example, `11` to be smaller than `2`.
* The level specified by `LevelName` should not be included in the lookup value table.

## Example

### Approximate threshold lookup

This example looks up a salary against an ordered threshold table and returns the corresponding tax value.

The lookup uses one node for the value to search, one node for the sorted index, and one node for the returned result.

**Input node:** `'Lookup Value'`

| Employee | Salary |
| -------- | ------ |
| E1       | 50.000 |

**Input node:** `'Index'`

| Sequence | Threshold |
| -------- | --------- |
| 0        | 0         |
| 1        | 10.000    |
| 2        | 30.000    |
| 3        | 60.000    |

**Input node:** `'Return Value'`

| Sequence | Tax   |
| -------- | ----- |
| 0        | 0     |
| 1        | 0     |
| 2        | 1.000 |
| 3        | 2.000 |
| 4        | 3.000 |

**Formula:** `LOOKUP('Lookup Value', 'Index', 'Return Value')`

| Employee | LOOKUP Result |
| -------- | ------------- |
| E1       | 1.000         |

## Expanding the sequence dimension

The sequence dimension is a default dimension that defines the steps of a time value. In rare cases, it can be helpful or even necessary to edit the sequence dimension to support complex scenarios on a monthly level.

Additional information on the sequence dimension and how to modify it can be found here: [Configure the Sequence Dimension](/configure-the-sequence-dimension/)

## Related Functions

| Function           | When to use instead                                                                                                             |
| ------------------ | ------------------------------------------------------------------------------------------------------------------------------- |
| [FILTER](/filter/) | When you want to filter rows by explicit criteria instead of mapping one value range to another through an ordered index table. |
| [IF](/if/)         | When you want explicit conditional branching logic instead of using an approximate lookup table.                                |