Filter syntax¶

Combining filters¶

Combination is done in the usual CSS way. A rule with two filters separated by a comma affects any features that match either filter, while a rule with two filters separated by only whitespace affects only features that match both filters. Here’s an example using a basic attribute filter (described below):

/* Matches places where the lake is flooding */
[rainfall>12] [lakes>1] {
    fill: black;
}

/* Matches wet places */
[rainfall>12], [lakes>1] {
    fill: blue;
}

When writing a selector that uses both and and or combinators, remember that the and combinator has higher precedence. For example:

restricted [cat='2'], [cat='3'], [cat='4'] [@sd <= 200k] [@sd > 100k] {
  fill: #EE0000;
}

The above selector should be read as:

• typename is ‘restricted’ and cat='2' or

• cat='3' or

• cat='4' and scale is between 100000 and 200000

If instead the intention was to combine in or just the three cat filters, the right syntax would have been:

restricted [cat='2' or cat='3' or cat='4'] [@sd <= 200k] [@sd > 100k] {
  fill: #EE0000;
}

Which should be read as:

• typename is ‘restricted’ and

• (cat='2' or cat='3' or cat='4') and

• scale is between 100000 and 200000

Filtering on data attributes¶

An attribute filter matches some attribute of the data (for example, a column in a database table). This is probably the most common type of filter. An attribute filter takes the form of an attribute name and a data value separated by some predicate operator (such as the less-than operator <).

Supported predicate operators include the following:

For example, to only render outlines for the states whose names start with letters in the first half of the alphabet, the rule would look like:

[STATE_NAME<='M'] {
    stroke: black;
}

Filtering on type¶

When dealing with data from multiple sources, it may be useful to provide rules that only affect one of those sources. This is done very simply; just specify the name of the layer as a filter:

states {
    stroke: black;
}

Filtering by ID¶

For layers that provide feature-level identifiers, you can style specific features simply by specifying the ID. This is done by prefixing the ID with a hash sign (#):

#states.2 {
    stroke: black;
}

Filtering by rendering context: scale¶

Often, there are aspects of a map that should change based on the context in which it is being viewed. For example, a road map might omit residential roads when being viewed at the state level, but feature them prominently at the neighborhood level. Details such as scale level are presented as pseudo-attributes; they look like property filters, but the property names start with an @ symbol:

[roadtype = 'Residential'][@sd > 100k] {
    stroke: black;
}

The context details that are provided are as follows:

The scale value can be expressed as a plain number, for for brevity and readability the suffixes k (kilo), M (mega), G (giga) can be used, for example:

[@sd > 100k]
[@sd < 12M]
[@sd < 1G]

Filtering by rendering context: zoom¶

Similar to scale based control, one can limit the application of a rule by referring to a zoom level using @z. By default the values of @z are interpreted as scale denominators in the WebMercatorQuad tile matrix set (the well known one used by most XYZ tile providers).

And example with results similar to the scale example would look as follows:

[roadtype = 'Residential'][@z < 13] {
    stroke: black;
}

The @z variable can be used with the following comparison operators: <, <=, =, >, >=. In case of equality, the range between the median scale denominator below the zoom level, and the one above, is used. So for example, when using @z=13, the translated SLD will use a range of scale denominators between 50k and 100k.

If other reference gridsets are desired, it’s possible to use the tileMatrixSet directive at the top of the CSS file. For example:

@tileMatrixSet 'EPSG:4326'

[roadtype = 'Residential'][@z < 13] {
    stroke: black;
}

The context details that are provided are as follows:

Filtering symbols¶

When using symbols to create graphics inline, you may want to apply some styling options to them. You can specify style attributes for built-in symbols by using a few special selectors:

For more discussion on using these selectors, see Styled marks.

Global rules¶

Sometimes it is useful to have a rule that matches all features, for example, to provide some default styling for your map (remember, by default nothing is rendered). This is accomplished using a single asterisk * in place of the usual filter. This catch-all rule can be used in complex expressions, which may be useful if you want a rule to provide defaults as well as overriding values for some features:

* {
    stroke: black;
}