QGIS Planet

./configure -prefix /opt/qt593 -debug -opensource -confirm-license -ccache -platform linux-clang

make -j9
sudo make install

/opt/qt593/bin/qmake WebKit.pro
make -j9
sudo make install

/opt/qt593/bin/qmake qscintilla.pro
make -j9
sudo make install

/opt/qt593/bin/qmake qwt.pro
make -j9
sudo make install

You may have seen my earlier tutorial where I described how to make nice looking hillshaded maps in QGIS using SRTM elevation data. Well, we don’t have to stop with just one hillshade layer on a map, it is possible to overlay multiple hillshades; a procedure that can increase the visual quality and detail. The following image is the hillshade we made before. Once you re-create a hillshade, following the previous tutorial, you can head to the next step (note that brightness and contrast settings may be different due to changes in how QGIS generates and displays hillshades).

We can improve the SRTM hillshade further by adding ESRI’s World Hillshade layer, which uses multi-directional illumination (also called a Swiss Hillshade in tribute to the celebrated Swiss cartographer Eduard Imhof). In addition, World Hillshade has a much higher resolution than SRTM 30m data in some regions of the world, it is 2m for most of the England and Wales, 10m for most of the US, 5m for Spain and 3m for Holland etc. The only drawback is that the style of this layer is somewhat controversial, some love it, some hate it, it looks like it’s illuminated from above, but mixing it with the SRTM hillshade obviates some of it criticised flaws.

To add the World Hillshade layer in QGIS go to the Layer Menu – Add Layer – Add ArcGIS MapServer Layer – click New and add the following URL:

https://services.arcgisonline.com/arcgis/rest/services/Elevation/World_Hillshade/MapServer

Notice QGIS 2.18 no longer needs a plugin to add ESRI layers, it new has this functionality built in. Also, open the url in a browser such as Firefox, it brings up a webpage that describes the layer. We also see links to other other layers. Yes, they can all be added to QGIS by simply taking the URL of the webpage that describe the layer and connecting to it via the ArcGIS MapServer Layer connector.

Name the layer World Hillshade and click Connect, then click and highlight the layer it connects to. Finally, click the Add button to add the layer to the canvas.

Next, we need to adjust the properties of the World Hillshade layer to properly overlay it above the SRTM hillshade layer. Make sure the World hillshade layer is the topmost layer. In the Layers Panel, right click Layer properties and in the window that opens up, click Style (if not visible). Next, change the Layer Blending mode (under color rendering) to Overlay. Adjust the layer’s brightness to around -20 and leave contrast at 0. If you find the scene is still too dark, brighten the SRTM Hillshade by increasing the layer’s brightness. You may also have to change (lower) the Min value of the Min – Max value boxes. Leave the contrast at 0 for the SRTM hillshade. Also, don’t brighten it too much as it might become washed out, loose detail, especially in bright areas. Play around the controls, settings may vary depending on the SRTM data you download and the version of QGIS you use.

Here’s a comparison in Ireland, a ring like structure of hills with a central peak. No, it’s not a meteorite crater. It’s a different kind of geological marvel, the Slieve Gullion Complex and its ring dyke; the deeply eroded remains of a 410 million year old Caledonian volcano. The SRTM hillshade is on the left and World Hillshade + SRTM hillshade is on the right (click on the image, it’s best appreciated full size):

We can see the World Hillshade + SRTM Hillshade layer shows much finer detail. We see a parallel array of roughly north-south orientated lines, these are fractures and faults that cut the Slieve Gullion Complex that were perhaps enhanced by glacial erosion. Also, look carefully, there seems to be some roads meandering across the landscape (hint, bottom of the map and right of the scale bar). You should get even better results with higher resolution World Hillshade data. We also notice that bending SRTM derived hillshade with World Hillshade adds a naturalistic illumination not apparent in multi-directional hillshading. So we have the best of both worlds, a high resolution hillshade and realistic looking illumination.

Hope you found this tutorial helpful.

References:

Baxter, S., 2008. A Geological Field Guide to Cooley Gullion, Mourne & Slieve Croob [pdf]. Geological Survey of Ireland, Dublin. p. 43-53.

Geologists display the dip and strike of rock layers on geological maps using a dip and strike symbol, where dip in degrees indicates the maximum angle a rock layer descends relative to the horizontal. However, it is not directly possible in QGIS 2.18, using basic label settings, to place and orient a dip label next to a dip and strike symbol.

However, there is a way around this issue using Trigonometry and editing the layer’s Attribute Table. This method may be useful for controlling the position and orientation of labels around point features in general. The first step involves adding values to the Attribute Table. First, add these two new columns:

• Angle – 0° is North and values increases clockwise up to 359°
• Distance – label distance from a point feature

You can add Angle and Distance values to these columns manually or use the Field Calculator (see below) to add values if you have lots of points. Also, I chose Map Units (not millimeters) for Symbol Size, Font Size and Distance for my map, as I prefered to keep symbol size, font size and position of labels fixed when zooming in and out.

Note – I use Strike (Angle) and Label Distance (Distance)  in my Attribute Table

The next step is to control the position of the label around the points using trigonometry. Right click the points layer and choose:

Properties – Labels – Placement

Check that Offset From Point is checked and then click the Data Defined Override next to the Offset X, Y boxes and choose Edit. The Expression String Builder will appear. Enter the following expression in the Expression String Builder window:

to_string ( ((-1) * ( “Distance” )) * cos ( radians ( “Angle” ))) ||’,’|| to_string (((-1) * ( “Distance” )) * sin ( radians ( “Angle” )) )

The expression takes the angle and distance values from the Attribute Table (edited earlier) and calculates an X, Y label position relative to the point feature. You may also optionally control the angle of a symbol or icon itself via:

Layer Properties – Style – click Data Defined Override icon – Edit

Then enter the following expression in the Data Defined Override dialogue:

“Angle” – 90

Finally, to control the rotation of label text, so text follows the orientation (angle) of a rotating symbol or icon, choose:

Layer Properties – Labels – Placement – Data Defined – Rotation

Click the Data Defined Override Icon again and then choose Edit. Enter the following expression in the Data Defined Override dialogue:

(“Angle” – 90) * -1

The following geological map of the Old Head of Kinsale in southern Ireland shows the results of the above procedure. We see that the dip labels rotate and currently follow the orientation of the dip and strike symbols (note that the points are at the intersection of the T symbol).

Geological Survey of Ireland – Creative Commons Attribution 4.0 license

You may have several different symbols, of various sizes, each requiring an appropriate label distance expressed in the Attribute Table. It took me a few tries before I found the right distances for my geological symbols, from 90 to 230 meters distance depending on the symbol size and type.

Lastly, the expressions “Angle” – 90 and (“Angle” – 90) * -1 were necessary in my case because I needed to place my labels next to the dip and strike symbol’s barb. You may need to use a different expression e.g. “Angle” and (“Angle”) * -1, or a value other than 90° depending on the symbol used and the prefered label placement location. Some trial and error is may be required to find the correct label position.

• fixed some bugs and feature requests on the official QGIS plugin repo that I’m maintaining since the very beginning
• make the QGIS official plugin repository website mobile-friendly
• QGIS Server Python Plugin API refactoring, I’ve completed the work on the new API, thanks to the ongoing server refactoring it’s now much cleaner than it was in the first version
• attribute table bugs: I started to address some nasty bugs in the attribute table, some of those were fixed during the week right after the hackfest
• unified add layer button, we had a productive meeting where we decided the path forward to implement this feature, thanks to Boundless that is funding the development, this feature is what’s I’m currently working on these days

Last week we kicked off the first (of hopefully many) Australian QGIS hackfests Developers Meetings. It was attended by 3 of the core QGIS development team: Nathan Woodrow, Martin Dobias and myself (Nyall Dawson), along with various family members. While there’s been QGIS hackfests in Europe for over 10 years, and others scattered throughout various countries (I think there was a Japanese one recently… but Twitter’s translate tool leaves me with little confidence about this!), there’s been no events like this in the Southern hemisphere yet. I’ve been to a couple in Europe and found them to be a great way to build involvement in the project, for both developers and non-developers alike.

In truth the Australian hackfest plans began mostly an excuse for Nathan and I to catch up with Martin Dobias before he heads back out of this hemisphere and returns to Europe. That said, Nathan and I have long spoken about ways we can build up the QGIS community in Australia, so in many ways this was a trial run for future events. It was based it in Noosa, QLD (and yes, we did manage to tear ourselves away from our screens long enough to visit the beach!).

Nathan Woodrow (@NathanW2), myself (@nyalldawson), and Martin Dobias (@wonder-sk)

Here’s a short summary of what we worked on during the hackfest:

• Martin implemented a new iterator style accessor for vertices within geometries. The current approach to accessing vertices in QGIS is far from optimal. You either have the choice of an inefficient methods (eg QgsGeometry.asPolyline(), asPolygon(), etc) which requires translations of all vertices to a different data structure (losing any z/m dimensional values in the process), or an equally inefficient QgsAbstractGeometry.coordinateSequence() method, which at least keeps z/m values but still requires expensive copies of every vertex in the geometry. For QGIS 3.0 we’ve made a huge focus on optimising geometry operations and vertex access is one of the largest performance killers remaining in the QGIS code. Martin’s work adds a proper iterator for the vertices within a geometry object, both avoiding all these expensive copies and also simplifying the API for plugins. When this work lands traversing the vertices will become as simple as

for v in geom.vertices():
   ... do something with the vertex!

• Martin is also planning on extending this work to allow simple iteration over the parts and rings within geometries too. When this lands in QGIS we can expect to see much faster geometry operations.
• Nathan fixed a long standing hassle with running standalone PyQGIS scripts outside of the QGIS application on Windows. In earlier versions there’s a LOT of batch file mangling and environment variable juggling required before you can safely import the qgis libraries within Python. Thanks to Nathan’s work, in QGIS 3.0 this will be as simple as just making sure that the QGIS python libraries are included in your Python path, and then importing qgis.core/gui etc will work without any need to create environment variables for OSGEO/GDAL/PLUGINS/etc. Anyone who has fought with this in the past will definitely appreciate this change, and users of Python IDEs will also appreciate how simple it is now to make the PyQGIS libraries available in these environments.
• Nathan also worked on “profiles” support for QGIS 3.0. This work will add isolated user profiles within QGIS, similar to how Chrome handles this. Each profile has it’s own separate set of settings, plugins, etc. This work is designed to benefit both plugin developers and QGIS users within enterprise environments. You can read more about what Nathan has planned for this here.
• I continued the ongoing work of moving long running interface “blocking” operations to background tasks. In QGIS 3.0 many of these tasks churn away in the background, allowing you to continue work while the operation completes. It’s been implemented so far for vector and raster layer saving, map exports to images/PDF (not composers unfortunately), and obtaining feature counts within legends. During the hackfest I moved the layer import which occurs when you drag and drop a layer to a destination in the browser to a background task.
• On the same topic, I took some inspiration from a commit in Sourcepole’s QGIS fork and reworked how composer maps are cached. One of my biggest gripes with QGIS’ composer is how slow it is to work with when you’ve got a complex map included. This change pushes the map redrawing into a background thread, so that these redraws no longer “lock up” the UI. It makes a HUGE difference in how usable composer is. This improvement also allowed me to remove those confusing map item “modes” (Cache/Render/Rectangle) – now everything is redrawn silently in the background whenever required.
• Lastly, I spent a lot of time on a fun feature I’ve long wanted in QGIS – a unified search “locator” bar. This feature is heavily inspired by Qt Creator’s locator bar. It sits away down in the status bar, and entering any text here fires up a bunch of background search tasks. Inbuilt searches include searching the layers within the current project (am I the only one who loses layers in the tree in complex projects!?), print layouts in the project, processing algorithms, and menu/toolbar actions. The intention here is that plugins will “take over” and add additional search functionality, such as OSM place names searching, data catalog searches, etc. I’m sure when QGIS 3.0 is released this will quickly become indispensable!

The upcoming QGIS 3.0 locator bar

Big thanks go out to Nathan’s wife, Stacey, who organized most of the event and without whom it probably would never have happened, and to Lutra Consulting who sponsored an awesome dinner for the attendees.

We’d love this to be the first of many. The mature European hackfests are attended by a huge swath of the community, including translators, documentation writers, and plugin developers (amongst others). If you’ve ever been interested in finding out how you can get more involved in the project it’s a great way to dive in and start contributing. There’s many QGIS users in this part of the world and we really want to encourage a community of contributors who “give back” to the project. So let Nathan or myself know if you’d be interested in attending other events like this, or helping to organize them locally yourself…

A lot of cartographers have a love/hate relationship with label halos. On one hand they can be an essential technique for improving label readability, especially against complex background layers. On the other hand they tend to dominate maps and draw unwanted attention to the map labels.

In this post I’m going to share my preferred techniques for using label halos. I personally find this technique is a good approach which minimises the negative effects of halos, while still providing a good boost to label readability. (I’m also going to share some related QGIS 3.0 news at the end of this post!)

Let’s start with some simple white labels over an aerial image:

These labels aren’t very effective. The complex background makes them hard to read, especially the “Winton Shire” label at the bottom of the image. A quick and nasty way to improve readability is to add a black halo around the labels:

Sure, it’s easy to read the labels now, but they stand out way too much and it’s difficult to see anything here except the labels!

We can improve this somewhat through a better choice of halo colour:

This is much better. We’ve got readable labels which aren’t too domineering. Unfortunately the halo effect is still very prominent, especially where the background image varies a lot. In this case it works well for the labels toward the middle of the map, but not so well for the labels at the top and bottom.

A good way to improve this is to take advantage of blending (or “composition”) modes (which QGIS has native support for). The white labels will be most readable when there’s a good contrast with the background map, i.e. when the background map is dark. That’s why we choose a halo colour which is darker than the text colour (or vice versa if you’ve got dark coloured labels). Unfortunately, by choosing the mid-toned brown colour to make the halos blend in more, we are actually lightening up parts of this background layer and both reducing the contrast with the label and also making the halo more visible. By using the “darken” blend mode, the brown halo will only be drawn for pixels were the brown is darker then the existing background. It will darken light areas of the image, but avoid lightening pixels which are already dark and providing good contrast. Here’s what this looks like:

The most noticeable differences are the labels shown above darker areas – the “Winton Shire” label at the bottom and the “Etheridge Shire” at the top. For both these labels the halo is almost imperceptible whilst still subtly doing it’s part to make the label readable. (If you had dark label text with a lighter halo color, you can use the “lighten” blend mode for the same result).

The only issue with this map is that the halo is still very obvious around “Shire” in “Richmond Shire” and “McKinlay” on the left of the map. This can be reduced by applying a light blur to the halo:

There’s almost no loss of readability by applying this blur, but it’s made those last prominent halos disappear into the map. At first glance you probably wouldn’t even notice that there’s any halos being used here. But if we compare back against the original map (which used no halos) we can see the huge difference in readability:

Compare especially the Winton Shire label at the bottom, and the Richmond Shire label in the middle. These are much clearer on our tweaked map versus the above image.

Now for the good news… when QGIS 3.0 is released you’ll no longer have to rely on an external illustration/editing application to get this effect with your maps. In fact, QGIS 3.0 is bringing native support for applying many types of live layer effects to label buffers and background shapes, including blur. This means it will be possible to reproduce this technique directly inside your GIS, no external editing or tweaking required!

It’s been a long time since I last blogged here. Let’s just blame that on the amount of changes going into QGIS 3.0 and move on…

One new feature which landed in QGIS 3.0 today is a processing algorithm for automatic coloring of a map in such a way that adjoining polygons are all assigned different color indexes. Astute readers may be aware that this was possible in earlier versions of QGIS through the use of either the (QGIS 1.x only!) Topocolor plugin, or the Coloring a map plugin (2.x).

What’s interesting about this new processing algorithm is that it introduces several refinements for cartographically optimising the coloring. The earlier plugins both operated by pure “graph” coloring techniques. What this means is that first a graph consisting of each set of adjoining features is generated. Then, based purely on this abstract graph, the coloring algorithms are applied to optimise the solution so that connected graph nodes are assigned different colors, whilst keeping the total number of colors required minimised.

The new QGIS algorithm works in a different way. Whilst the first step is still calculating the graph of adjoining features (now super-fast due to use of spatial indexes and prepared geometry intersection tests!), the colors for the graph are assigned while considering the spatial arrangement of all features. It’s gone from a purely abstract mathematical solution to a context-sensitive cartographic solution.

The “Topological coloring” processing algorithm

Let’s explore the differences. First up, the algorithm has an option for the “minimum distance between features”. It’s often the case that features aren’t really touching, but are instead just very close to each other. Even though they aren’t touching, we still don’t want these features to be assigned the same color. This option allows you to control the minimum distance which two features can be to each other before they can be assigned the same color.

The biggest change comes in the “balancing” techniques available in the new algorithm. By default, the algorithm now tries to assign colors in such a way that the total number of features assigned each color is equalised. This avoids having a color which is only assigned to a couple of features in a large dataset, resulting in an odd looking map coloration.

Balancing color assignment by count – notice how each class has a (almost!) equal count

Another available balancing technique is to balance the color assignment by total area. This technique assigns colors so that the total area of the features assigned to each color is balanced. This mode can be useful to help avoid large features resulting in one of the colors appearing more dominant on a colored map.

Balancing assignment by area – note how only one large feature is assigned the red color

The final technique, and my personal preference, is to balance colors by distance between colors. This mode will assign colors in order to maximize the distance between features of the same color. Maximising the distance helps to create a more uniform distribution of colors across a map, and avoids certain colors clustering in a particular area of the map. It’s my preference as it creates a really nice balanced map – at a glance the colors look “randomly” assigned with no discernible pattern to the arrangement.

Balancing colors by distance

As these examples show, considering the geographic arrangement of features while coloring allows us to optimise the assigned colors for cartographic output.

The other nice thing about having this feature implemented as a processing algorithm is that unlike standalone plugins, processing algorithms can be incorporated as just one step of a larger model (and also reused by other plugins!).

QGIS 3.0 has tons of great new features, speed boosts and stability bumps. This is just a tiny taste of the handy new features which will be available when 3.0 is released!

A year ago I have asked QGIS’s community what were their favourite QGIS new features from 2015 and published this blog post. This year I decided to ask it again. In 2016, we add the release of the second long-term release (2.14 LTR), and two other stable versions (2.16 and 2.18).

2016 was again very productive year for the QGIS community, with lots of improvements and new features landing on QGIS source code, not to speak of all the work already in place for QGIS 3. This is a great assurance of the project’s vitality.

As a balance, I have asked users to choose wich were their favorite new features during 2016 (from the visual changelogs list). As a result, I got the following Top 5 features list.

5 – Paste a style to multiple selected layers or to all layers in a legend group (2.14)

This is a productivity functionaly that I just realized that existed now, with so many people voting on it. If copy/paste styles was, in my opinion, a killer feature, being able to use it in multiple layers or even a group is just great.

4 – fTools plugin has been replaced with Processing algorithms (2.16)

While checking the Vector Menu, the tools seem the same as previous version, but it’s when you open them that you understand the difference. All vector tools, provided until now by the fTools core plugin, were replaced by equivalent processing Algoritms. For the users it means easier access to more functionality, like running the tools in batch mode, or getting outputs as temporary layers. Besides some of the tools have been improved.

3 – Virtual layers (2.14)

This is definitly one of my favourite new features, and it seems I’m not alone. With virtual layers you can run SQL queries using the layers loaded in the project, even when the layers are not stored in a relational database. We are not talking about WHERE statments to filter data, with this you can do real SQL queries, with spatial analysis, aggregations, and so on. Besides, virtual layers will act as VIEWs and any changes to any of the input layers will automatically update the layer.

2 – Speed and memory improvements (2.14)

It’s no surprise that speed and memory improvements we one of the most voted features. Lots of improvements were made for loading and managing large datasets, and this have a tremendous impact in all users. According to the changelog, zoom is faster, selecting features is faster, updating attributes on selected features is faster, and it consumes less memory. So don’t be afraid to put QGIS to the test.

1 – Trace digitising tool (2.14)

If you do lots of digitising, you better look into this new feaure that landed on QGIS 2.14. It allows you to digitize new feature by using other layers boundaries. Besides the quality improvement of layers topology, this can make digitizing almost feel pleasing and fast! Just click the first point, move your mouse around other features edged to pick up more vertex.

There were other new features that also made the delight of many users. For example, several improvements on the labeling, Georeference outputs (eg PDF) from composer (2.16), Filter legend by expression (2.14), 2.5D Renderer. Personally, the Style docker made my day/year. But you can check the full results of the survey, if you like.

Obviously, this list means nothing at all. All new features were of tremendous value, and will be useful for thousands (yes thousands) of people. It was a mere exercise as, with such a diverse QGIS crowd, it would be impossible to build a list that would fit us all. Besides, there were many great enhancements, introduced during 2016, that might have fallen under the radar for most users. Check the visual changelogs for a full list of new features.

On my behalf, to all developers, sponsors, and general QGIS contributors, once again

THANK YOU VERY MUCH FOR YOUR TREMENDOUS WORK!

I wish you a fantastic 2017.

I’ve recently spent some time optimising the performance of various QGIS plugins and algorithms, and I’ve noticed that there’s a few common performance traps which developers fall into when fetching features from a vector layer. In this post I’m going to explore these traps, what makes them slow, and how to avoid them.

As a bit of background, features are fetched from a vector layer in QGIS using a QgsFeatureRequest object. Common use is something like this:

request = QgsFeatureRequest()
for feature in vector_layer.getFeatures(request):
    # do something

This code would iterate over all the features in layer. Filtering the features is done by tweaking the QgsFeatureRequest, such as:

request = QgsFeatureRequest().setFilterFid(1001)
feature_1001 = next(vector_layer.getFeatures(request))

In this case calling getFeatures(request) just returns the single feature with an ID of 1001 (which is why we shortcut and use next(…) here instead of iterating over the results).

Now, here’s the trap: calling getFeatures is expensive. If you call it on a vector layer, QGIS will be required to setup an new connection to the data store (the layer provider), create some query to return data, and parse each result as it is returned from the provider. This can be slow, especially if you’re working with some type of remote layer, such as a PostGIS table over a VPN connection. This brings us to our first trap:

Trap #1: Minimise the calls to getFeatures()

A common task in PyQGIS code is to take a list of feature IDs and then request those features from the layer. A see a lot of older code which does this using something like:

for id in some_list_of_feature_ids:
    request = QgsFeatureRequest().setFilterFid(id)
    feature = next(vector_layer.getFeatures(request))
    # do something with the feature

Why is this a bad idea? Well, remember that every time you call getFeatures() QGIS needs to do a whole bunch of things before it can start giving you the matching features. In this case, the code is calling getFeatures() once for every feature ID in the list. So if the list had 100 features, that means QGIS is having to create a connection to the data source, set up and prepare a query to match a single feature, wait for the provider to process that, and then finally parse the single feature result. That’s a lot of wasted processing!

If the code is rewritten to take the call to getFeatures() outside of the loop, then the result is:

request = QgsFeatureRequest().setFilterFids(some_list_of_feature_ids)
for feature in vector_layer.getFeatures(request):
    # do something with the feature

Now there’s just a single call to getFeatures() here. QGIS optimises this request by using a single connection to the data source, preparing the query just once, and fetching the results in appropriately sized batches. The difference is huge, especially if you’re dealing with a large number of features.

Trap #2: Use QgsFeatureRequest filters appropriately

Here’s another common mistake I see in PyQGIS code. I often see this one when an author is trying to do something with all the selected features in a layer:

for feature in vector_layer.getFeatures():
    if not feature.id() in vector_layer.selectedFeaturesIds():
        continue

    # do something with the feature

What’s happening here is that the code is iterating over all the features in the layer, and then skipping over any which aren’t in the list of selected features. See the problem here? This code iterates over EVERY feature in the layer. If you’re layer has 10 million features, we are fetching every one of these from the data source, going through all the work of parsing it into a QGIS feature, and then promptly discarding it if it’s not in our list of selected features. It’s very inefficient, especially if fetching features is slow (such as when connecting to a remote database source).

Instead, this code should use the setFilterFids() method for QgsFeatureRequest:

request = QgsFeatureRequest().setFilterFids(vector_layer.selectedFeaturesIds())
for feature in vector_layer.getFeatures(request):
    # do something with the feature

Now, QGIS will only fetch features from the provider with matching feature IDs from the list. Instead of fetching and processing every feature in the layer, only the actual selected features will be fetched. It’s not uncommon to see operations which previously took many minutes (or hours!) drop down to a few seconds after applying this fix.

Another variant of this trap uses expressions to test the returned features:

filter_expression = QgsExpression('my_field > 20')
for feature in vector_layer.getFeatures():
    if not filter_expression.evaluate(feature):
        continue

    # do something with the feature

Again, this code is fetching every single feature from the layer and then discarding it if it doesn’t match the “my_field > 20” filter expression. By rewriting this to:

request = QgsFeatureRequest().setFilterExpression('my_field > 20')
for feature in vector_layer.getFeatures(request):
    # do something with the feature

we hand over the bulk of the filtering to the data source itself. Recent QGIS versions intelligently translate the filter into a format which can be applied directly at the provider, meaning that any relevant indexes and other optimisations can be applied by the provider itself. In this case the rewritten code means that ONLY the features matching the ‘my_field > 20’ criteria are fetched from the provider – there’s no time wasted messing around with features we don’t need.

Trap #3: Only request values you need

The last trap I often see is that more values are requested from the layer then are actually required. Let’s take the code:

my_sum = 0
for feature in vector_layer.getFeatures(request):
    my_sum += feature['value']

In this case there’s no way we can optimise the filters applied, since we need to process every feature in the layer. But – this code is still inefficient. By default QGIS will fetch all the details for a feature from the provider. This includes all attribute values and the feature’s geometry. That’s a lot of processing – QGIS needs to transform the values from their original format into a format usable by QGIS, and the feature’s geometry needs to be parsed from it’s original type and rebuilt as a QgsGeometry object. In our sample code above we aren’t doing anything with the geometry, and we are only using a single attribute from the layer. By calling setFlags( QgsFeatureRequest.NoGeometry ) and setSubsetOfAttributes() we can tell QGIS that we don’t need the geometry, and we only require a single attribute’s value:

my_sum = 0
request = QgsFeatureRequest().setFlags(QgsFeatureRequest.NoGeometry).setSubsetOfAttributes(['value'], vector_layer.fields() )
for feature in vector_layer.getFeatures(request):
    my_sum += feature['value']

None of the unnecessary geometry parsing will occur, and only the ‘value’ attribute will be fetched and populated in the features. This cuts down both on the processing required AND the amount of data transfer between the layer’s provider and QGIS. It’s a significant improvement if you’re dealing with larger layers.

Conclusion

Optimising your feature requests is one of the easiest ways to speed up your PyQGIS script! It’s worth spending some time looking over all your uses of getFeatures() to see whether you can cut down on what you’re requesting – the results can often be mind blowing!

The 6th QGIS UK user group meeting in Scotland is happening on the 3rd November 2016.  It is being hosted by the EDINA University of Edinburgh at the Informatics Forum and is sponsored by thinkWhere, Ordnance Survey, Angus Council and Cawdor Forestry.  Tickets are available through Eventbrite.

The almost final programme of presentations and lightning talks is as follows:

• Phil Taylor (CEH) – How deep is your loch?
• Fiona Hemsley-Flint – QGIS server
• University of Edinburgh – packaging and deploying QGIS
• Anouk Lang – Mapping narrative: QGIS in the humanities classroom
• Art Lembo (Salisbury University, MD) – terrain analysis with massively parallel processing techniques (embarrasingly so)
• Neil Benny (thinkWhere) – finding the heart of Scotland / viewshed analysis
• Tom Chadwin – qgis2web and coding a QGIS plugin
• Pete Wells (Lutra) – WMTS previews and XYZ support
• Stephen Bathgate – decision support system in Forestry
• Tim Manners (Ordnance Survey) – Creating an indoor routable network with QGIS and pgRouting
• Andrew Whitelee – QGIS in forestry/ecology
• Ross McDonald (Angus Council) – Them thar hills: shaded, textured and blended
• Michal Michalski (The Origins of Doha and Qatar Project) – DOHA: Doha Online Historical Atlas
• eeGeo – Using QGIS to create 3D indoor maps

Doors open from 9:00. Registration shortly thereafter. Start and welcome at 9:45 and a planned finish at 16:30. Geobeers to follow.