Guide

Shapefile (SHP): The Geospatial Industry's Most-Used Vector Format

PC By Pablo Cirre • Updated May 7, 2026

What Is a Shapefile?

The Shapefile is the most widely deployed vector geospatial data format in the world — and it is not a single file. Introduced by ESRI in 1994 alongside ArcView GIS 2, a Shapefile is actually a bundle of files sharing the same base name but different extensions, all of which must travel together for the data to be usable.

Despite its age and well-documented limitations, the Shapefile remains the lingua franca of geospatial data exchange. Government agencies, OpenStreetMap exports, Census Bureau TIGER files, natural resource agencies, and transportation departments all distribute data in Shapefile format. Understanding it is unavoidable for anyone working with GIS.

The File Bundle

A complete Shapefile requires at minimum three files:

Extension	Role	Format
`.shp`	Geometry storage — coordinates of features	Binary
`.dbf`	Attribute table — rows of data for each feature	dBASE IV
`.shx`	Spatial index — byte offsets into the .shp file	Binary

And commonly includes:

Extension	Role
`.prj`	Coordinate Reference System as WKT string
`.cpg`	Code page declaration for .dbf text encoding (e.g., UTF-8)
`.qpj`	QGIS extended PRJ with more complete CRS metadata
`.sbn` / `.sbx`	Spatial index for ArcGIS performance
`.xml`	ISO 19139 metadata

If you receive a .shp file without its .dbf companion, you have geometry but no attribute data. If you are missing the .prj, software will ask you to specify the coordinate system manually — a common source of projection mismatch errors.

Geometry Types

Shapefile supports exactly one geometry type per file:

Type Code	Geometry	Example Use
1	Point	City locations, sample sites, GPS waypoints
3	Polyline (MultiLineString)	Roads, rivers, pipelines, contours
5	Polygon (MultiPolygon)	Countries, parcels, land use zones
8	MultiPoint	Point clouds, earthquake epicenters
11	PointZ	3D points with elevation
13	PolylineZ	3D paths — tunnels, flight tracks
15	PolygonZ	3D surfaces — building footprints with height
21	PointM	Measured points — milepost markers

A single Shapefile cannot mix geometry types — all roads must go in one file, all polygons in another. This is a fundamental limitation that trips up new GIS users.

The DBF Attribute Table

The .dbf (dBASE IV) file stores attribute data for each geometry:

Maximum 10 characters per field name — the single most complained-about Shapefile limitation
Maximum 255 bytes per text field value
Maximum 2,147,483,647 records (2 GB file size ceiling)
No native support for Unicode in the original spec (.cpg file partially addresses this)
No support for NULL values in integer/float fields (stored as 0, which is ambiguous)
No date-time type — only date (YYYYMMDD)

These constraints mean attributes like "CountyName" become "CNTYNM" and "PopulationDensityPerSquareKilometer" becomes "PDPSKM" — completely opaque abbreviations that require a separate data dictionary to interpret.

The .prj Projection File

The projection file contains a Well-Known Text (WKT) string describing the Coordinate Reference System:

GEOGCS["GCS_WGS_1984",
    DATUM["D_WGS_1984",
        SPHEROID["WGS_1984",6378137.0,298.257223563]],
    PRIMEM["Greenwich",0.0],
    UNIT["Degree",0.0174532925199433]]

Without this file, software must guess the CRS — often defaulting to WGS84. When the actual data is in a projected system (UTM, State Plane, ETRS89), features will appear in the wrong location or at the wrong scale. Always distribute the .prj file alongside your Shapefile.

Reading Shapefiles in Python

GeoPandas (the standard library for vector GIS in Python):

import geopandas as gpd

# Read a Shapefile
gdf = gpd.read_file("countries.shp")

print(gdf.crs)               # coordinate reference system
print(gdf.dtypes)            # column types
print(gdf.head())            # first few rows including geometry

# Spatial query — select features that intersect a bounding box
from shapely.geometry import box
bbox = box(-10, 35, 30, 70)   # roughly Europe
europe = gdf[gdf.intersects(bbox)]

# Reproject to Web Mercator for tiled mapping
gdf_mercator = gdf.to_crs(epsg=3857)

# Write back as Shapefile
gdf_mercator.to_file("countries_mercator.shp")

# Write as GeoJSON instead
gdf_mercator.to_file("countries.geojson", driver="GeoJSON")

Fiona (lower-level, closer to OGR):

import fiona

with fiona.open("countries.shp") as src:
    print(src.crs)
    print(src.schema)        # geometry type + field types
    for feature in src:
        geom = feature["geometry"]
        props = feature["properties"]
        print(props["NAME"], geom["type"])

Converting Shapefiles with ogr2ogr

# Shapefile → GeoJSON
ogr2ogr -f GeoJSON countries.geojson countries.shp

# Shapefile → GeoPackage (modern replacement)
ogr2ogr -f GPKG countries.gpkg countries.shp

# Shapefile → PostGIS
ogr2ogr -f "PostgreSQL" PG:"host=localhost dbname=gis user=postgres" \
        countries.shp -nln countries -overwrite

# Reproject on the fly (from EPSG:4326 to EPSG:3857)
ogr2ogr -f "ESRI Shapefile" countries_mercator.shp countries.shp \
        -s_srs EPSG:4326 -t_srs EPSG:3857

# Filter by attribute
ogr2ogr -f GeoJSON europe.geojson countries.shp -where "CONTINENT='Europe'"

# Clip to bounding box
ogr2ogr -f "ESRI Shapefile" clipped.shp input.shp \
        -clipsrc -10 35 30 70

Why Move to GeoPackage?

GeoPackage (.gpkg) is the OGC-standardized SQLite-based replacement for Shapefile:

Feature	Shapefile	GeoPackage
Single file	No (bundle)	Yes (one .gpkg)
Max field name	10 characters	1000+ characters
Max file size	2 GB	140 TB
Multiple layers	One per file set	Multiple in one file
Unicode support	Partial (.cpg)	Full UTF-8
Geometry types	Single type per file	Mixed per layer
NULL values	Ambiguous (0)	True NULL
Date-time support	Date only	Full datetime
Raster support	No	Yes (tiles)

QGIS, ArcGIS Pro, GDAL, and PostGIS all support GeoPackage natively. The US National Geospatial-Intelligence Agency (NGA) mandated GeoPackage as the preferred format for geospatial data exchange in 2021. For new projects, GeoPackage is almost always the better choice.

Working with the US Census TIGER Shapefiles

The US Census Bureau distributes TIGER/Line Shapefiles — the authoritative dataset for US administrative boundaries, roads, and demographic geography:

# Download 2023 counties Shapefile
wget https://www2.census.gov/geo/tiger/TIGER2023/COUNTY/tl_2023_us_county.zip
unzip tl_2023_us_county.zip

# Inspect with ogrinfo
ogrinfo -al -so tl_2023_us_county.shp

# Convert to GeoJSON, keeping only a few fields
ogr2ogr -f GeoJSON counties.geojson tl_2023_us_county.shp \
        -select "NAME,STATEFP,COUNTYFP,ALAND,AWATER"

The TIGER files use EPSG:4269 (NAD83) — not WGS84. The difference is sub-meter for most applications, but for precision surveying or alignment with GPS data, you may need to reproject.

Shapefile Size Limits in Practice

The 2 GB limit applies separately to the .shp and .dbf files. A Shapefile with complex polygon geometries (coastlines, watershed boundaries) can exhaust the 2 GB .shp limit without having many features. The OpenStreetMap full-planet Shapefile exports routinely exceed this limit and must be split by geographic region or converted to GeoPackage or PostGIS for practical use.

Shapefile cannot represent topology — shared boundaries between adjacent polygons are stored as duplicate coordinate sequences, leading to slivers and gaps after independent editing of adjacent features. Formats like PostGIS enforce topology explicitly; Shapefile never has and never will.

Related conversions

Frequent conversions across the catalogue:

Frequently Asked Questions

The Shapefile format was designed in 1994 as a multi-file system: .shp stores binary geometry records, .dbf stores attribute data in dBASE format, .shx stores a byte-offset index, and .prj stores the coordinate system. These were originally separate concerns mapped to separate file types. All files must stay together — zip them together when sharing. This is the most common beginner mistake: emailing just the .shp and wondering why it fails to load.

On KaijuConverter every file is processed inside an isolated container, encrypted in transit (TLS 1.3) and at rest, and automatically deleted after 60 minutes with multi-pass overwrite. We never train on, share, or analyze user content. For maximum privacy on extremely sensitive material, prefer offline tools (ImageMagick, FFmpeg, LibreOffice) that you control end-to-end.

The .dbf (dBASE IV) format used for Shapefile attributes limits column names to 10 characters. This was a constraint of the dBASE file format from the early 1980s, long before ESRI adopted it. It means long descriptive names like "PopulationDensity" must be truncated to "POPDENSITY" or abbreviated to "POP_DEN". This limitation is one of the primary reasons GIS professionals are migrating to GeoPackage, which supports field names up to 1000+ characters.

The .dbf (dBASE IV) formato used para Shapefile attributes limits column names to 10 characters. This was a constraint of the dBASE formatoo de arquivo de the early 1980s, long antes ESRI adopted it. It means long descriptive names like "PopulationDensity" must be truncated to "POPDENSITY" ou abbreviated to "POP_DEN". This limitation is one of the primary reasons GIS profissionals are migrating to GeoPackage, which suporta field names up to 1000+ characters.

The .dbf (dBASE IV) Format used für Shapefile attributes limits column names to 10 characters. This was a constraint des dBASE Dateiformat von the early 1980s, long vor ESRI adopted it. It means long descriptive names like "PopulationDensity" must be truncated to "POPDENSITY" oder abbreviated to "POP_DEN". This limitation is one des primary reasons GIS professionells are migrating to GeoPackage, which unterstützt field names up to 1000+ characters.

The .dbf (dBASE IV) formato used para Shapefile attributes limits column names to 10 characters. This was a constraint del dBASE formatoo de archivo de the early 1980s, long antes ESRI adopted it. It means long descriptive names like "PopulationDensity" must be truncated to "POPDENSITY" o abbreviated to "POP_DEN". This limitation is one del primary reasons GIS profesionals are migrating to GeoPackage, which soporta field names up to 1000+ characters.

For 95% of use cases, yes — server-side ImageMagick, FFmpeg and LibreOffice produce identical output to the same tools on your laptop. Desktop software wins for: extremely large files (multi-GB), batch jobs of thousands of files, scripted pipelines, or content too sensitive to upload. KaijuConverter caps at 25 MB per file on the free tier (up to 2 GB on paid plans).

GeoPackage (GPKG) is the modern, OGC-standardized replacement. It solves nearly all Shapefile limitations: it's a single SQLite file (no bundle), supports field names up to 1000 characters, handles files up to 140 TB, stores multiple layers in one file, supports true NULL values, full Unicode, and mixed geometry types. All major GIS software supports both formats. The US NGA mandated GeoPackage as the preferred format in 2021. For any new project, choose GeoPackage unless you specifically need Shapefile compatibility.

GeoPackage (GPKG) is the moderNão, OGC-standardized replacement. It solves nearly all Shapefile limitations: it's a single SQLite arquivo (no bundle), suporta field names up to 1000 characters, handles arquivos up to 140 TB, stores multiple layers in one file, suporta true NULL values, full Unicode, e mixed geometry types. All major GIS software suporta both formatoos. The US NGA mandated GeoPackage como o preferred formato in 2021. para any new project, choose GeoPackage unless you specifically need Shapefile compatibilidade.

GeoPackage (GPKG) is the modern, OGC-standardized replacement. It solves nearly all Shapefile limitations: it's a single SQLite Datei (no bundle), unterstützt field names up to 1000 characters, handles Dateien up to 140 TB, stores multiple layers in one file, unterstützt true NULL values, full Unicode, und mixed geometry types. All major GIS Software unterstützt both Formate. The US NGA mandated GeoPackage als das preferred Format in 2021. für any new project, choose GeoPackage unless you specifically need Shapefile Kompatibilität.

GeoPackage (GPKG) is the moderNo, OGC-standardized replacement. It solves nearly all Shapefile limitations: it's a single SQLite archivo (no bundle), soporta field names up to 1000 characters, handles archivos up to 140 TB, stores multiple layers in one file, soporta true NULL values, full Unicode, y mixed geometry types. All major GIS software soporta both formatoos. The US NGA mandated GeoPackage como el preferred formato in 2021. para any new project, choose GeoPackage unless you specifically need Shapefile compatibilidad.

Most format conversions are lossy by design — JPG, MP3, MP4, WebP all discard perceptual data to save bytes. Going through a lossy intermediate compounds the loss. To minimize visible/audible drift: convert from the original master, choose a higher quality setting, and avoid converting back and forth between lossy formats.

Use GeoPandas: pip install geopandas, then gdf = gpd.read_file("myfile.shp"). GeoPandas wraps Fiona (which wraps GDAL/OGR) and returns a GeoDataFrame — a pandas DataFrame with a geometry column. You get full spatial operations (intersect, buffer, dissolve, sjoin), CRS management, and easy export to any format GDAL supports. For lightweight access without spatial analysis, pyshp (shapefile) is a pure-Python alternative with no compiled dependencies.

Use GeoPandas: pip install geopandas, then gdf = gpd.read_file("myfile.shp"). GeoPandas wraps Fiona (which wraps GDAL/OGR) e returns a GeoDataFrame — a pandas DataFrame com a geometry column. You get full spatial operations (intersect, buffer, dissolve, sjoin), CRS management, e fácil export to any formato GDAL supports. para lightweight access sem spatial analysis, pyshp (shapefile) is a pure-Python alternative com no compiled dependencies.

Use GeoPandas: pip install geopandas, then gdf = gpd.read_file("myfile.shp"). GeoPandas wraps Fiona (which wraps GDAL/OGR) und returns a GeoDataFrame — a pandas DataFrame mit a geometry column. You get full spatial operations (intersect, buffer, dissolve, sjoin), CRS management, und einfach export to any Format GDAL supports. für lightweight access ohne spatial analysis, pyshp (shapefile) is a pure-Python alternative mit no compiled dependencies.

Use GeoPandas: pip install geopandas, then gdf = gpd.read_file("myfile.shp"). GeoPandas wraps Fiona (which wraps GDAL/OGR) y returns a GeoDataFrame — a pandas DataFrame con a geometry column. You get full spatial operations (intersect, buffer, dissolve, sjoin), CRS management, y fácil export to any formato GDAL supports. para lightweight access sin spatial analysis, pyshp (shapefile) is a pure-Python alternative con no compiled dependencies.

Yes — KaijuConverter accepts multiple files in a single drop and returns a ZIP. For very large batches (thousands of files) consider command-line tools or our API: <code>find . -name "*.heic" -exec magick {} {.}.jpg \;</code> or similar one-liners scale to millions of files when run locally.