Converting from extended axis bin number to non-extended bin number, FindBin() vs FindFixBin()

[Fr4nk]

Hi All,

I am curious to see if there is a quick way to convert from the extended axis bin number, such as the one that would be obtained when using the FindBin() method to the non-extended bin number that would be obtained from FindFixBin().

Thanks!

-Frank

ROOT Version: 6.22/06
Platform: MacOS Big Sur
Compiler: gcc

---

[moneta]

Hi,

There is no different bin number, FindBin(x), if the axis is extendable and if x is larger than the axis maximum, will extend the axis and then return the corresponding bin number. FindFixBin(x) will instead, for the same x case, return as bin number, the overflow number (nbins+1).
I am not sure if this answers to your question.

Best,

Lorenzo

[Fr4nk]

@moneta I think I previously misunderstood the difference between the two methods, FindBin() and FindFixBin().

As a test I have done the following:

I generated a histogram:

TH2F *h_p = new TH2F("h_p", "Element array", 20, -100, 100, 20, -100, 100);

Which represents a grid over an area which I would like to fit sets of (x, y) points on and obtain which box/bin a given point is in. For instance, I will have an event at some point within the grid, within the range of (-100, 100) along both axes, and I would like to use:

element = h_p->FindBin(x, y, 0);

to determine the bin that contains the point. For my check I have done the following to make sure I return the same points:

x_test = h_p->GetXaxis()->GetBinCenter(element);
y_test = h_p->GetYaxis()->GetBinCenter(element);

cout << "test position = (" << x_test << " ," << y_test << ")" << endl;
cout << "seeded position = (" << x << " ," << y << ")" << endl;

upon compilation I obtain the following:

test position = (1905 ,1905)
seeded position = (-72.269 ,-10.6774)

test position = (1025 ,1025)
seeded position = (-77.7099 ,-53.9449)

test position = (835 ,835)
seeded position = (-44.8618 ,-66.9224)

test position = (2575 ,2575)
seeded position = (-61.6721 ,11.6775)

test position = (1375 ,1375)
seeded position = (53.6068 ,-41.5784)

test position = (3565 ,3565)
seeded position = (49.3616 ,56.9929)

test position = (2275 ,2275)
seeded position = (74.2571 ,-3.31977)

test position = (1505 ,1505)
seeded position = (-39.6788 ,-32.3741)

test position = (2765 ,2765)
seeded position = (-90.0729 ,24.7655)

test position = (2045 ,2045)
seeded position = (62.5263 ,-13.7962)

While I understand that the bin center and the seeded point used to determine the bin, will not exactly match, I expected them to be within a range of at maximum 5 units of each other.

I assume my understanding is still flawed. Any and all help is greatly appreciated.

[moneta]

Hello,
What you are doing is not correct, because TH1::FindBin returns a global bin number for the 2-dim histogram and you are using it as an axis bin number.
You have two possibilities:

1. get separately the axis bin number for the x axis and y axis:

xbin = h_p->GetXaxis()->FindBin(x); 
x_test = h_p->GetXaxis()->GetBinCenter(xbin); 
ybin = h_p->GetYaxis()->FindBin(y); 
y_test = h_p->GetYaxis()->GetBinCenter(ybin);

2. Convert from the global bin number to the axis bin number:

element = h_p->FindBin(x, y, 0);
int xbin, ybin, bin = 0;
h_p->GetBinXYZ(element, xbin, ybin,zbin); 

and then use xbin and ybin as above.

Best

Lorenzo

[Fr4nk]

@moneta Thank you so much, that fixed that issue! I have another question I am hoping you can answer. Do you happen to know how the global bin number is assigned to each bin? For example is it like axis bin: (xbin, ybin) = (0, 0) :: global bin = 0, and (xbin, ybin) = (20, 20) :: global bin = 400 ?

[moneta]

See ROOT: hist/hist/src/TH3.cxx Source File
where the global bin is computed from the bin axis number

Lorenzo