GeoTiffMetadata - GeoTIFF File Metadata

GeoTiffMetadata is a Spark data source that reads GeoTIFF file metadata without decoding pixel data, similar to gdalinfo. It returns one row per file with metadata including dimensions, coordinate system, band information, tiling, overviews, and compression.

This is useful for:

• Cataloging and inventorying large collections of raster files
• Detecting Cloud Optimized GeoTIFFs (COGs) by checking tiling and overview status
• Inspecting file properties before loading full raster data
• Building spatial indexes over raster file collections

COG detection

Cloud Optimized GeoTIFFs (COGs) are GeoTIFF files with internal tiling and overviews optimized for cloud access. The geotiff.metadata data source reports these properties directly:

df = sedona.read.format("geotiff.metadata").load("/path/to/rasters/")
cogs = df.filter("isTiled AND size(overviews) > 0")
cogs.select("path", "compression", "overviews").show(truncate=False)

Read GeoTIFF metadata

ScalaJavaPython

val df = sedona.read.format("geotiff.metadata").load("/path/to/rasters/")
df.show()

Dataset<Row> df = sedona.read().format("geotiff.metadata").load("/path/to/rasters/");
df.show();

df = sedona.read.format("geotiff.metadata").load("/path/to/rasters/")
df.show()

You can also use glob patterns:

df = sedona.read.format("geotiff.metadata").load("/path/to/rasters/*.tif")

Or load a single file:

df = sedona.read.format("geotiff.metadata").load("/path/to/image.tiff")

Output schema

Each row represents one GeoTIFF file with the following columns:

Column	Type	Description
path	String	File path
driver	String	Format driver ("GTiff")
fileSize	Long	File size in bytes
width	Int	Image width in pixels
height	Int	Image height in pixels
numBands	Int	Number of bands
srid	Int	EPSG code (0 if unknown)
crs	String	Coordinate Reference System as WKT
geoTransform	Struct	Affine transform parameters
cornerCoordinates	Struct	Bounding box
bands	Array[Struct]	Per-band metadata
overviews	Array[Struct]	Overview (pyramid) levels
metadata	Map[String, String]	File-wide TIFF metadata tags
isTiled	Boolean	Whether the file uses internal tiling
compression	String	Compression type (e.g., "LZW", "Deflate")

geoTransform struct

Field	Type	Description
upperLeftX	Double	Origin X in world coordinates
upperLeftY	Double	Origin Y in world coordinates
scaleX	Double	Pixel size in X direction
scaleY	Double	Pixel size in Y direction
skewX	Double	Rotation/shear in X
skewY	Double	Rotation/shear in Y

cornerCoordinates struct

Field	Type	Description
minX	Double	Minimum X (west)
minY	Double	Minimum Y (south)
maxX	Double	Maximum X (east)
maxY	Double	Maximum Y (north)

bands array element

Field	Type	Description
band	Int	Band number (1-indexed)
dataType	String	Data type (e.g., "REAL_32BITS")
colorInterpretation	String	Color interpretation (e.g., "Gray", "Red")
noDataValue	Double	NoData value (null if not set)
blockWidth	Int	Internal tile/block width
blockHeight	Int	Internal tile/block height
description	String	Band description
unit	String	Unit type (e.g., "meters")

overviews array element

Field	Type	Description
level	Int	Overview level (1, 2, 3, ...)
width	Int	Overview width in pixels
height	Int	Overview height in pixels

Examples

Inspect band information

df = sedona.read.format("geotiff.metadata").load("/path/to/image.tif")
df.selectExpr("path", "explode(bands) as band").selectExpr(
    "path",
    "band.band",
    "band.dataType",
    "band.noDataValue",
    "band.blockWidth",
    "band.blockHeight",
).show()

Filter by spatial extent

df = sedona.read.format("geotiff.metadata").load("/path/to/rasters/")
df.filter("cornerCoordinates.minX > -120 AND cornerCoordinates.maxX < -100").select(
    "path", "width", "height", "srid"
).show()

Get overview details

df = sedona.read.format("geotiff.metadata").load("/path/to/image.tif")
df.selectExpr("path", "explode(overviews) as ovr").selectExpr(
    "path", "ovr.level", "ovr.width", "ovr.height"
).show()

Select specific columns

Select only the columns you need:

df = (
    sedona.read.format("geotiff.metadata")
    .load("/path/to/rasters/")
    .select("path", "width", "height", "numBands")
)
df.show()