Common GIS Vector Operations Explained

GIS vector operations change, combine, divide, select, or construct geographic features. Clip can trim features to a study area, Intersect can retain only shared geography, Buffer can create zones around features, Dissolve can remove boundaries between related features, and Simplify can reduce geometric detail.

Those operations may all appear together in a GIS toolbox, but they do fundamentally different kinds of work. Some compare two layers. Some transform one layer. Some reorganise features without changing their overall coverage. Others construct entirely new geometry.

The most useful way to understand vector operations is therefore not to memorise a long catalogue of tool names. It is to start with the geographic change you need to make and identify the family of operations that expresses it.

What is a vector operation in GIS?

Vector data represents geographic features using points, lines, and polygons. A vector operation takes those geometries, their relationships, their attributes, or some combination of them and produces a new spatial result.

Consider a few different questions:

A planner has a national road dataset but needs only roads inside a project boundary.

A conservation analyst wants to know which protected areas overlap proposed infrastructure.

A researcher has many district polygons and wants one polygon for each region.

A transport analyst wants to identify places within 500 metres of a road.

A cartographer has extremely detailed boundaries that need to render more efficiently at small map scales.

All five tasks involve vector data, but the desired transformation is different in each case. The first suggests Clip, the second an overlay such as Intersect, the third Dissolve, the fourth Buffer, and the fifth some form of Simplification.

That relationship between the question and the transformation is the central idea behind vector processing.

Common vector operations at a glance

• If you need to…: Keep features only inside an area; Common operation: Clip; What it broadly does: Trims input geometry using another boundary

• If you need to…: Keep only geography shared by layers; Common operation: Intersect; What it broadly does: Constructs their overlapping geography

• If you need to…: Preserve the complete overlay of two polygon layers; Common operation: Union; What it broadly does: Keeps overlapping and non-overlapping areas divided by both sets of boundaries

• If you need to…: Remove areas covered by another layer; Common operation: Difference / Erase; What it broadly does: Subtracts one geometry from another

• If you need to…: Create an area around a feature; Common operation: Buffer; What it broadly does: Constructs a polygon within a specified distance

• If you need to…: Combine adjacent or related features; Common operation: Dissolve; What it broadly does: Merges geometry, often according to a shared attribute

• If you need to…: Put multiple compatible datasets into one layer; Common operation: Merge / Append; What it broadly does: Combines their features rather than spatially overlaying them

• If you need to…: Separate multipart geometry into individual features; Common operation: Explode; What it broadly does: Turns parts of a multipart feature into separate records

• If you need to…: Reduce geometric detail; Common operation: Simplify; What it broadly does: Removes vertices according to a simplification rule

• If you need to…: Add geometric detail between vertices; Common operation: Densify; What it broadly does: Inserts additional vertices along geometry

• If you need to…: Enclose a collection of geometry; Common operation: Convex or concave hull; What it broadly does: Constructs a boundary around input features

• If you need to…: Partition space around input points; Common operation: Voronoi; What it broadly does: Creates regions associated with the nearest input point

• If you need to…: Create triangles from input vertices; Common operation: Delaunay triangulation; What it broadly does: Constructs a connected triangular network

• If you need to…: Group features and summarise their attributes; Common operation: Aggregate; What it broadly does: Produces grouped geometry and calculated attributes

These descriptions are deliberately broad. Particular GIS packages can use slightly different names, expose different options, or separate one conceptual operation into several tools. The question being answered is usually more stable than the interface terminology.

Overlay operations compare geography

Some of the most familiar vector operations belong to the overlay family. They use the spatial relationship between two layers to determine which geometry should survive.

Clip, Intersect, Union, Difference, and Symmetrical Difference all belong here, but each preserves a different part of the inputs.

Imagine two partially overlapping polygon layers, A and B.

Clip can retain the part of A that lies inside B. Intersect retains the geography shared by A and B. Difference can retain the part of A that does not overlap B. Union, in its common polygon-overlay sense, preserves the areas covered by either layer and subdivides them according to boundaries from both.

Symmetrical Difference keeps the parts belonging to A or B but excludes their shared area.

These operations can sometimes create deceptively similar maps. For example, clipping A with B and intersecting A with B may produce the same visible geometry. Their analytical meaning still differs because Clip treats one input primarily as a boundary, while Intersect treats the relationship between the inputs as part of the result.

Clip vs Intersect vs Union examines that distinction in detail.

Overlay is useful when the geometry of a spatial relationship matters. If you merely need to know which polygon contains a point or which features satisfy a relationship without creating new geometry, a spatial selection or spatial join may be more appropriate.

Difference and Erase answer a subtraction question

Difference operations are another form of overlay, but the question is easier to understand as subtraction:

What remains of A after the parts covered by B are removed?

A land parcel intersected by a restricted zone, for example, could be processed with Difference to retain only the unrestricted portion.

Some GIS software uses names such as Difference and Erase for closely related operations. The precise terminology and attribute behaviour can vary, so the important conceptual distinction is between keeping the overlap and removing it.

This is the counterpart to operations such as Intersect. Intersect asks what two geometries share; Difference asks what remains when their shared portion is removed from one of them.

Dissolve removes boundaries within a layer

Not every vector operation compares two datasets.

Dissolve usually works within a single layer. It combines features that belong together, removing boundaries that no longer matter for the intended representation.

Suppose a country dataset contains district polygons and each district has a region attribute. Dissolving by region can create one feature for each region by combining the geometry of districts with the same regional value.

Nothing has been clipped to a new study area and there is no second layer defining an overlap. The operation is reorganising existing coverage according to a grouping rule.

This is why Dissolve should not be confused with Union simply because both can cause polygon boundaries to disappear. A union operation concerns the combined spatial extent of geometry; Dissolve is commonly about grouping features and removing internal boundaries according to how the data should be represented.

The distinction becomes especially important when attributes are involved, because dissolving several records into one raises a second question: what should happen to their non-geometric values?

A dedicated treatment of What Does Dissolve Do? can address those grouping and attribute decisions more fully.

Merge and Append combine datasets, not their spatial relationships

Another common confusion arises between Merge, Append, and spatial overlay.

Suppose you have one road layer for the northern half of a country and another compatible road layer for the southern half. If the goal is simply to place all those road features into one dataset, you generally do not need Intersect or Union.

You need a data-combination operation such as Merge or Append.

The geometries do not need to overlap, touch, or interact spatially. Their features are being brought together because they belong in the same output dataset.

Conceptually, Merge and Append operate more like combining collections of records, whereas overlay operations construct results from spatial relationships.

The exact distinction between Merge and Append also depends partly on the software and storage workflow: one tool may create a new output layer from several inputs, while another adds features to an existing destination. That implementation-specific question belongs in the narrower Merge vs Append vs Union article rather than being resolved by this overview.

Buffer turns distance into an area

A Buffer constructs an area around a point, line, or polygon.

A 500-metre buffer around a school can represent all locations within that geometric distance of the school. A buffer around a road can create a corridor. A positive buffer around a polygon expands its footprint, while GIS tools may also support negative polygon buffers that shrink the geometry inward.

The simplicity of the picture can hide an important analytical issue: a buffer represents geometric proximity, not necessarily real-world accessibility or influence.

A hospital may be 2 kilometres away in straight-line distance but much farther by road. A river, railway, mountain, border crossing, or disconnected street network can make a nearby place difficult to reach.

The appropriate operation therefore depends on what within a distance actually means in the question. Buffer is appropriate when geometric distance is the intended model. Travel-time accessibility requires network analysis instead.

Buffer calculations are also sensitive to how distance is represented in the coordinate reference system, which is why the choice of CRS and units matters when creating distance-based geometry.

Explode and multipart operations change feature structure

A geometry and a feature are not always the same thing.

One feature can contain several disconnected geometric parts. A country containing islands, for example, might be represented as one MultiPolygon feature rather than one feature for every land mass.

An Explode operation can separate those parts into individual features. Conversely, other operations can collect separate geometries into multipart results.

This changes the structure of the dataset even when the visible geographic coverage remains similar.

That distinction becomes important in later processing. Counting features, assigning attributes, calculating statistics, selecting individual pieces, or exporting to another system can produce different results depending on whether disconnected geometry is stored as one multipart feature or several independent features.

This is why Explode vs Dissolve is a useful distinction: Explode separates geometry structure, while Dissolve commonly combines features according to a grouping rule.

Simplification changes geometric detail

Real geographic features can contain thousands or millions of vertices. That detail may be useful at one scale and unnecessary at another.

Geometry simplification reduces the number of vertices while attempting to retain the significant form of a line or polygon.

A highly detailed coastline, for example, might be appropriate for local analysis but excessive for a small-scale world map. Simplifying it can reduce storage, processing, and rendering demands.

Simplification is not merely file compression, however. It changes geometry.

Different algorithms make different decisions about which vertices can be removed, and the chosen tolerance determines how much detail is sacrificed. Excessive simplification can move boundaries, distort narrow features, remove small shapes, or create unsuitable geometry for subsequent analysis.

There is therefore no universal simplification level that is simply “better”. The correct amount depends on why the geometry is being simplified and at what scale or analytical precision it must remain useful.

Smoothing is related but conceptually different. Simplification primarily reduces detail; smoothing primarily changes the form of lines to make them less angular or irregular. A result can look visually simpler without being the product of a simplification operation.

Densification does the opposite kind of geometric change

Where simplification removes vertices, Densification inserts them.

Why add points to a line that already represents the same path?

Because some transformations and calculations operate on the segments between vertices. A line represented by only a few widely spaced vertices may behave differently during reprojection, curved approximation, measurement, or subsequent geometric processing than a line containing intermediate points.

Densification therefore increases the explicit sampling of a geometry without necessarily changing its intended path.

More vertices do not automatically make geographic data more accurate. They simply provide a denser geometric representation of the shape being processed.

Hulls create enclosing geometry

Some operations construct a new boundary around existing features.

The best-known example is the convex hull: the smallest convex geometry capable of enclosing the input. A common visual analogy is a rubber band stretched around a group of points.

A convex hull cannot bend inward between points. If a tighter enclosing shape is needed, a concave hull can follow the distribution more closely, depending on the algorithm and parameters being used.

Hulls are useful when the question concerns an enclosing footprint rather than the exact geometry of the inputs. They can provide a simple representation of spatial extent, but they should not automatically be interpreted as observed boundaries. A hull connecting wildlife observations, for example, is a constructed geometric envelope; it is not by itself evidence that the entire enclosed area represents habitat or species presence.

That distinction between constructed and observed geography is important whenever a new geometry is inferred from sparse inputs.

Voronoi and Delaunay construct spatial relationships

Other operations create geometry from the relative arrangement of points.

A Voronoi diagram partitions space so that each region is associated with the input point nearest to locations inside that region. If the points represent facilities, each Voronoi cell describes the area that is geometrically closer to its facility than to the others.

A Delaunay triangulation connects input points into triangles according to their spatial arrangement and is mathematically related to the Voronoi construction.

These operations are useful geometric structures, but their outputs should not be confused with real service areas, administrative territories, or observed relationships. A Voronoi cell around a hospital, for example, describes nearest-point geometry. It does not account for road networks, capacity, jurisdiction, physical barriers, or patient behaviour.

As with Buffer, the operation answers precisely the geometric question it was designed to answer—not every real-world question that happens to sound similar.

Aggregation changes the level at which features are represented

Aggregation groups observations or features into a coarser representation.

The grouping may be spatial, categorical, or both. Individual records can become totals by district; detailed polygons can become regional features; point observations can be summarised into grid cells.

This is partly a geometry operation and partly a data operation because the resulting attributes need rules as well. Numeric values might be summed or averaged, categories might need counts, and some attributes may have no sensible aggregate at all.

Disaggregation moves in the opposite conceptual direction, attempting to distribute information from a coarser unit into finer spatial units. Unlike simply exploding multipart geometry, however, disaggregation often requires assumptions about how values should be allocated.

Areal interpolation is one family of methods for transferring values between incompatible polygon systems. Those topics require substantially more reasoning than a mechanical geometry transformation, so they deserve dedicated treatment rather than being presented as simple opposites of Dissolve.

Rasterisation and vectorisation cross the data-model boundary

Not every operation in a vector workflow produces another vector layer.

Rasterisation converts vector information into cells. Vectorisation creates vector geometry from raster information or another gridded representation.

These operations change the data model itself.

That means decisions about cell size, geometry construction, classification, boundaries, and attribute transfer can affect what information survives the conversion. Converting a polygon into a raster and then vectorising it again should not be assumed to reproduce the original geometry exactly.

For that reason, rasterisation and vectorisation sit at the edge of the vector-operations family. Understanding them properly also requires understanding the difference between vector objects and raster surfaces.

The same visible result can come from different operations

One reason vector processing becomes confusing is that two different workflows can occasionally produce maps that appear similar.

Clip and Intersect can yield the same visible polygon in a simple case. Dissolve and geometric union can both make internal boundaries disappear. Merge can put features from two layers into one output, while an overlay Union can also produce a single output layer from two inputs.

The difference appears when you ask more precise questions:

• Question: Is another layer defining a boundary or contributing information?; What it helps distinguish: Clip vs Intersect

• Question: Do I want to keep non-overlapping areas?; What it helps distinguish: Intersect vs Union

• Question: Am I subtracting overlap?; What it helps distinguish: Difference vs Intersect

• Question: Am I combining records or combining geometry?; What it helps distinguish: Merge vs Dissolve

• Question: Should internal boundaries disappear?; What it helps distinguish: Merge vs Dissolve

• Question: Am I modelling geometric distance or travel through a network?; What it helps distinguish: Buffer vs accessibility analysis

• Question: Am I reducing vertices or changing the apparent shape?; What it helps distinguish: Simplification vs smoothing

• Question: Am I separating multipart geometry or grouping separate features?; What it helps distinguish: Explode vs Dissolve

This is why the operation should follow from the intended meaning of the output rather than from the icon that seems to produce the desired picture.

Before running a vector operation, define the output

A useful habit is to describe the intended output in plain language before choosing the GIS tool.

Instead of beginning with “Should I use Intersect?”, begin with:

I need the portions of each parcel that fall inside each flood zone, and I need the parcel identifier and flood classification on every resulting piece.

That description makes Intersect a natural candidate.

Or:

I have separate district datasets with the same structure and need all of their features in one layer without changing their geometry.

That points towards Merge or Append.

Or:

I need one regional feature for every set of districts sharing the same region code.

That points towards Dissolve or an aggregation workflow.

This approach also makes it easier to recognise when a familiar tool is inappropriate. If the desired output cannot be described clearly, selecting an operation first is likely to hide rather than resolve the ambiguity.

Vector operations change the meaning of the data

The most important thing to remember about vector processing is that it does not merely rearrange shapes on a map.

An overlay can turn one feature into many. A Dissolve can turn many records into one. A Buffer creates geography that did not exist as an input feature. A Simplification changes the coordinates used to represent a boundary. A hull infers an enclosing geometry from other geometries. Aggregation changes the geographic level at which information is expressed.

Those transformations may be exactly what the analysis requires, but the output is now derived data.

Its meaning depends on the operation, its inputs, and its parameters.

Understanding common vector operations is therefore less about memorising a toolbox than learning to ask three questions:

What geography should survive? What new geometry should be created? What should each output feature mean?

Once those are clear, Clip, Intersect, Union, Dissolve, Buffer, Simplify, and the rest stop looking like an arbitrary list of GIS commands. They become different ways of expressing different geographic transformations.

References

1. QGIS Documentation — Vector overlay. Documents common overlay operations including Clip, Difference, Intersection, Symmetrical Difference, and Union, and describes which parts of the input geometries each operation retains.

2. QGIS Documentation — Vector geometry. Documents geometry-processing operations including Buffer, Dissolve, Aggregate, Densify, Simplify, Convex Hull, Delaunay Triangulation, and geometry restructuring tools.

3. QGIS Documentation — Vector general. Documents general vector data-management operations including Merge Vector Layers and helps distinguish dataset combination from spatial overlay.