ST_Reclassify Raster Pixel Remapping Reference - PolarDB

Reclassifies pixel values in a raster object by mapping original values to new values based on a classification expression.

The returned raster has the same spatial reference and resolution as the input. The number of output bands is determined by the reclassexpr parameter.

Syntax

raster ST_Reclassify(raster raster_obj,
                     cstring reclassexpr DEFAULT NULL,
                     cstring storageOption DEFAULT '')

Parameters

Parameter	Description
`raster_obj`	The raster object to reclassify.
`reclassexpr`	A JSON string array that defines the classification rules for each band.
`storageOption`	A JSON string that configures the storage format of the returned raster.

reclassexpr

Each element in the reclassexpr array is a JSON object that specifies reclassification rules for one band.

Parameter	Type	Default	Description
`band`	integer	—	The band index to reclassify. Band indexes start at 0.
`remap`	object	—	The pixel value mapping rules. See remap.
`nodata`	boolean	`false`	Specifies whether to treat nodata pixels as nodata in the output. If `true`, pixels with the nodata value remain nodata. If `false`, non-nodata values are calculated as regular numeric values.
`nodataValue`	float8	`0`	The nodata value to assign in the output.

remap

The remap object maps original pixel value ranges to new pixel values. Each key-value pair defines one mapping rule.

Key format — The key specifies the input pixel value range using interval notation:

Symbol	Meaning
`(`	Greater than (open left boundary)
`)`	Less than (open right boundary)
`[`	Greater than or equal to (closed left boundary)
`]`	Less than or equal to (closed right boundary)

The default boundary type is (] (open left, closed right).

Use commas to separate boundary values within the key. For example, (0,100,200] defines two intervals: (0, 100] and (100, 200].

Value format — The value specifies the output pixel values, separated by commas.

Three mapping methods are supported:

Method	Description	Example
Range-to-range	Input and output have the same number of values. Each input value maps directly to the corresponding output value.	`"300,400,500":"80,90,100"`
Range-to-value	Input has one more value than output. Each interval between consecutive input values maps to one output value.	`"(300,400,500]":"80,90"`
Value-to-value	Both input and output have exactly one value. A single pixel value maps to a single new value.	`"10":"1"`

Constraints:

Pixels outside all defined ranges are treated as nodata.
Ranges cannot overlap (no pixel value can belong to more than one range).

remap examples

Example 1 — Single band, range-to-value mapping

Reclassify band 0:

(0, 100] → 20
(100, 200] → 50
All other values → nodata

[
  {
    "band": 0,
    "remap": {
      "(0,100,200]": "20,50"
    }
  }
]

Example 2 — Multiple ranges

Pixels in (200, 300] and other gaps outside defined ranges are set to nodata.

[
  {
    "band": 0,
    "remap": {
      "(0,100,200]": "20,50",
      "(300,400,500]": "80,90,100"
    }
  }
]

Example 3 — Multi-band with nodata settings

Band 0:

(0, 100] → 20
(100, 200] → 50
All other values → 999 (set as nodata)

Band 1:

(400, 600] → linear interpolation from 20 to 90
(600, 800] → linear interpolation from 90 to 130
All other values → 0

[
  {
    "band": 0,
    "remap": { "(0,100,200]": "20,50" },
    "nodata": true,
    "nodataValue": 999
  },
  {
    "band": 1,
    "remap": { "(400,600,800]": "20,90,130" },
    "nodata": false,
    "nodataValue": 0
  }
]

storageOption

Parameter	Type	Default	Description
`chunking`	boolean	Same as input	Specifies whether to store the raster as chunks.
`chunkdim`	string	Same as input	The chunk dimensions. Takes effect only when `chunking` is `true`.
`chunktable`	string	`''`	The name of the chunk table. If left blank (`''`), a temporary chunk table with a random name is created and is valid only for the current session. Specify a name to retain an accessible raster object after the session ends.
`compression`	string	Same as input	The compression algorithm. Valid values: `none`, `jpeg`, `zlib`, `png`, `lzo`, `lz4`.
`quality`	integer	Same as input	The compression quality. Takes effect only for `jpeg` compression.
`interleaving`	string	Same as input	The interleaving method. Valid values: `bip` (band interleaved by pixel, BIP), `bil` (band interleaved by line, BIL), `bsq` (band sequential, BSQ).
`endian`	string	Same as input	The byte order. Valid values: `NDR` (little endian), `XDR` (big endian).
`celltype`	string	Same as input	The pixel type of the output raster.

Examples

-- Create a permanent table to store results.
CREATE TABLE rast_reclassify_result(id integer, rast raster);

-- Create a temporary table to store results.
CREATE TEMP TABLE rast_reclassify_result_temp(id integer, rast raster);

-- Reclassify band 0 and store the result in the temporary table.
INSERT INTO rast_reclassify_result_temp(id, rast)
SELECT 1, ST_Reclassify(rast, '[{"band":0,"remap":{"(0,100,200]":"20,50"}}]')
FROM reclass_table;