Group by: interval of 0.1 (decimal)

[rob6]

Hi all - just a quick check to see if I’m going mad, or have missed something.

In short, when using grouped by: is it possible to have a range that starts from a minus number and ends in a positive, with a decimal as the increment?

---

I have a set of data which are decimals (they represent percentages), so 0.7 = 70% etc.
I’ll be using the data for a graph but that doesn’t impact this part.

I’m using grouped by: to bucket the percentages, and count the number that sit within each.

Example raw data
-0.23
-0.21
0
0.2
0.1

Desired bucket count output:
-0.2 = 2
-0.1 = 0
0 = 1
0.1 = 1
0.2 = 1

This works as expected if I start the buckets from 0 and set the interval to 0.1 (But does not include negative values).

Or if I start from -1, I get no data if the interval is 0.1. But if I set it to 1, I do get the data expected. But for the percentage data it’s not that useful.

So it feels like it’s something to do with a negative starting value and decimal increment combined. Although a search and reference reading hasn’t given me anything concrete.

Here’s a shot of the data:

And here’s a shot of my expression:

And the grouping detail:

And here’s the output - just in a text box for easy debugging for now.

Thanks for taking the time to read, and if you have any ideas, I’ll be incredibly grateful!
Rob

[cmarchan]

Perhaps transforming the range from numbers to a text will allow you to group them without the negative number problem. Do not know if this affects other functionality you have that requires numbers.

[rob6]

Ah, a very good suggestion but as I’ll be using the data in a graph it requires numbers annoyingly.

Thanks for your thoughts though

[boston85719]

Try setting the type of grouping to exact instead of bucket…not sure if that will impact the way you want to group, or if it will allow for a negative starting value and a 0.1 interval, but my first guess as to how to get it to work.

[rob6]

Thanks! I tried this and a few other methods, but in the end I just went further up the chain and amended the data import process so that it comes in as 10 vs 0.1.