 Hello, this is Hans van der Kras senior lecturer at IT Delft Institute for Water Education. In this video I'm going to show how to use lookup tables with PC Raster in QGIS. Let's say we want to select only the houses class out of the buildg Raster here, then we can simply use the raster calculator, we don't need lookup tables, and this also to demonstrate that you can use the PC Raster maps with the raster calculator without any problem. So here I say when the buildg Raster equals value 1, which is for houses, then give me boolean true and otherwise give me boolean false. Now in the raster calculator it doesn't work if you use PC Raster as an output format, so stick to the geotiff. Then when I run it I get here a boolean layer, a geotiff with the houses. When I style it this becomes clear, so two classes. Let's add the label, zero is no houses and one is houses, and make the colors more intuitive because it's boolean. Let's make false white, so it looks like a background, and make true green. Now let's have a look at our road map. It has different types of roads, dirt road and tarmac, and I would like to reclassify it in boolean with road and no road, and that is very easy using a lookup table that you can simply make as an ASCII file in, for example, notepad. PC Raster uses a table in ASCII format where the first column is the original value from the raster, and the second is what it will be assigned, so in this case zero remains zero, one will be boolean true, and two will also be boolean true. I save the table as reclassroads. So I use a lookup operation, input is roads, then I load the lookup table, reclassroads, and I need to specify the output data type, in our case that's boolean, and I can save it, and let's call it isroad. Run it, and it creates a boolean layer which has cells which are no road and cells that are road. We can style it, we palette the unique values, we classify no road and the road, and use boolean red for false and green for true. But can we also classify ranges, because that would be very easy. Well, that's possible. Let's classify the build g layer into classes that are industry related and that are not industry related. So if I look at the legend, then I can see that classes zero, one, and two are no industry, and three, four, and five are industry related. So I can make a lookup table with ranges. Also that's a simple ASCII table. You can use the less than simple to say that it does not include that value of the range, and if we keep it open, it means all the values and we separate with a comma. So this means all the values less than two, including two, become boolean zero. And the second line means all the values from three, including three, until the highest value that it finds becomes boolean one, true. So I save it as reclass industry, and then I can use lookup, give the table the output again will be boolean, and I call it is industry, and I can run it and copy the style to make it a boolean look. There it is industry versus no industry. Just need to change the labels, but I'm not going to do that here. Set the projection for the question marks. Now we have only looked at boolean outputs, but what if we look at other types of outputs from reclassifications and those ranges. So let's reclassify our DTM layer. And what I'm going to do is specify three different ranges, everything smaller than 200 and 10 meters, not including 210 will become class one. Then from 210 and including 210 to 250, not including 250 becomes class two, and 250, including the 250, until the largest value that it finds becomes class number three. So that's an ordinal scale. Let's call it reclass DTM, run lookup, use the DTM as input and our new lookup table, reclass DTM. And the output is ordinal in this case. These are classes, but the order does matter, so then it becomes ordinal. Let's call it DTM ordinal. I run it, move it to the top, and now it needs proper styling. Let's also set the projection. Ordinal also classes, but then we need to use a ramp because the hierarchy of the order should be visible. So something like this. There's also a lookup linear tool, which interpolates values between the values that it finds in the table. If you click on the link, it will go to the page in the documentation of PC Raster. Here you see how it works. And here, for example, we have a table with input value 0257, and that will be an output 0110. And here we see an input of 1, which does not exist in the table, but then it takes the middle value, so it will be 0.5 in the result. So that's basically how it works with an interpolation. These were the lookup operations. I'm still working on making it possible to have multiple rasters as an input and use tables with multiple columns because that's also supported by PC Raster. But that's not ready yet. When it's ready, I'll post a video.