Toolbox Reference

01 What the Toolbox is

87 point-and-click geoprocessing tools across seven categories — and every one of them writes R code you can keep.

Open the Toolbox from the Toolbox button in the top toolbar. Pick a tool, fill in the form, and press Run. The result lands in your R environment as a new object and opens in the Data Viewer.

What makes this different from a normal GIS toolbox is the Code Preview pane at the bottom of every tool form. It shows the exact R that your form settings produce, live, as you type. Running the tool runs precisely that code in your session — there is no hidden step, and nothing happens that you can’t see, copy, or edit.

02 The Toolbox dialog

The dialog is split down the middle: a searchable tool tree on the left, and the selected tool’s form on the right.

03 The tool form

Every tool form is generated from the tool’s own definition, so they all behave the same way. Fields appear top to bottom, the code preview sits underneath, and the action buttons are pinned to the bottom-right.

Code Preview

The R your settings currently produce. It updates on every keystroke and never lags behind the form. This is the single most useful part of the Toolbox: it is how you learn the R equivalent of a GUI operation you already know.

Use only current view

When a dataset has an active selection or the map is zoomed to a subset, this checkbox restricts the tool to those features rather than the whole layer. The generated code slices the dataset first, so the restriction is visible in the preview like everything else.

Buttons

04 Field types

The parameter columns in the reference tables below use these field types.

The condition builder

Filter Rows builds compound conditions without you writing the expression. Press Add Condition for each clause and join clauses with AND or OR. The right-hand side can be a literal value (val) or another column (col).

05 Linearize and Calculation CRS

Two options that only appear on spatial tools, and that decide whether your curved geometry survives the operation.

Linearize curved geometry before calculation

Every tool in Spatial — sf and Spatial — geos carries this checkbox — the one exception is Points from XY Columns, which has no geometry to linearize. It is off by default, and off is what you usually want.

• Unticked — curved geometry (CIRCULARSTRING, COMPOUNDCURVE, CURVEPOLYGON) is passed through as-is, so sfarc can compute on the true arc and return an exact answer.
• Ticked — the input is chopped into straight segments first by wrapping it in ourgeo_linearize(). You lose exactness, but you gain access to operations GEOS can only perform on polylines.

Calculation CRS

Exact curved maths is planar. On geographic (lat/lon) data an arc has no constant radius, so nine tools — Buffer, Intersection, Union (both), Difference, Erase, Clip, Centroid and Simplify — add a Calculation CRS dropdown. The generated code projects into that CRS, computes, and projects the result back to the layer’s own CRS, so what you get out is still in the coordinate system you started in.

Raster tools have neither option. Linearize and Calculation CRS are sf/geos concepts; terra handles reprojection explicitly through Reproject Raster.

06 Piping tools together

Pressing Pipe → instead of Run chains tools into one R pipeline rather than creating an intermediate object per step. Configure a tool, press Pipe →, pick the next tool, and repeat; press Run on the last one to execute the whole chain at once.

The result is a single readable expression:

# Filter, then buffer, then dissolve — one pipeline, one output object
sites_final <- sites |>
  dplyr::filter(status == "active") |>
  sf::st_buffer(units::set_units(250, m)) |>
  sf::st_union()

• The output name is locked from the first step — the chain produces one object, not one per tool.
• Column dropdowns in later steps show the columns that will exist at that point, predicted from the earlier steps. Renaming, selecting or mutating a column in step one is reflected in step two’s dropdown without R having run yet.
• The banner across the top of the form traces the chain as you build it.
• Clear pipeline (the × on the banner) abandons the chain.

07 Tool reference

All 87 tools, by category.

The Parameters column lists each tool’s inputs in the order the form shows them; the output-name field is on every tool and is omitted here. The R call column names the main function the tool generates — useful when you want to look something up with ? in the console. Linearize and Calc CRS mark the tools that carry those extra options, described in section 05.

Filtering is optional — with the box empty every tool stays in the page, so your browser’s own find (Ctrl + F) searches all of them.

08 Table Operations 11 tools

Row- and column-level work on any dataset — tabular or spatial. Every tool here emits dplyr, so an sf layer keeps its geometry column throughout.

09 Joins & Combine 2 tools

Stack or splice two datasets together. For joining by location rather than by value, see Spatial — sf’s Spatial Join.

10 Aggregate & Reshape 5 tools

Collapse many rows into summary rows, and pivot between long and wide layouts.

11 Type & String 8 tools

Fix the things that are wrong with data as it arrives — wrong column types, padded whitespace, NAs, values that need splitting out of one field.

12 Spatial — sf 16 tools

Geometry operations through the sf package. These are the tools that understand curves: leave Linearize unticked and a CIRCULARSTRING stays exact.

13 Spatial — geos 11 tools

The same family of overlay and predicate operations routed through the geos package instead. Faster on large planar datasets, but GEOS has no concept of a true arc.

14 Spatial — Raster 34 tools

Gridded data through terra — reprojection, algebra, terrain, and the crossings between raster and vector.

Processing — 18 tools

Bands — 4 tools

Raster ↔ Vector — 7 tools

Conversion & Output — 5 tools