I have something that might help,

my original code snippet with the issues, described in my question, looked like this:

```
//loading a **root** file with **tree** name **Board** that has a double variable named **t_day** that contains the **time** info in **days**, it has another variable named **adc**
t1 = 0; //days
t2 = 31; //days
int tBin = (t2-t1)*24*60/10; //every 10 minutes (units of days)
double dt = 10./(60.*24.); //bin width (10 minutes in days)
TProfile *prof = new TProfile("prof","prof", tBin, t1, t2);
for(int i = 0; i < tBin; i++)
{
Time = t1 + i*dt;
TH1F *h = new TH1F("h", "h", 100, 0, 100);
Board->Draw("adc>>h", Form("t_day >= %g && t_day < %g", Time, Time + dt));
//I fit this histogram "h" and extract a parameter say "mean_h" and fill it to the TProfile "prof"
prof->Fill(Time + dt/2., mean_h);
}
prof->Draw("e1");
```

Then I changed the variable “**Time**” that I was feeding to the **TProfile**, instead of filling `Time + dt/2.`

I filled just `Time`

and my problem was gone! Nice **TProfile** with separated points

**The questions naturally arises:**

I am checking if the time variable of the root file, that is **t_day** is within my external loop time window, which is **Time = t1 + i * dt** and **Time + 10 minutes**, then using that piece of data. But while filling, of course I want to use the mean value of this time: **Time + dt/2.** to fill my **TProfile** **prof**, why can’t I do that? Why the lower **edge** that is **Time** works and gives me nice **TProfile** with separated points but becomes **strange histogram looking bar diagram** when I fill the mid value **Time + dt/2.** ?

**My Guess:**

Suppose my definition was:

`TProfile *prof = new TProfile("prof","prof", 5, 0, 5);`

In such a case, by definition, on the **X-axis** I have time bins:

`0, 1, 2, 3, 4, 5`

,

but I am filling the values

`0.5, 1.5, 2.5, 3.5, 4.5`

instead.

If **ROOT Tprofiles** was saving values at the middle of a bin then there shouldn’t have been any problem, but my guess is that it saves values at the beginning of the bin and not in the middle, therefore **0.5** was allocated to the value **0**, **1.5** to **1** and so on… can this **root** behavior be the **root cause** of my problem?

**VERY IMPORTANT EDIT**

When the **TProfile** looked like **TH1F** with **bars**, it was more **correct**, in the sense that each bins had right values and there were as many bars (with different values) as there were number of entries, but when I changed the fill variable from **Time + dt/2.** to just **Time**, bar problem got solved and I got nice **TProfile** with separated points, but some of the points got dumped in single bins, as a result, there were less **points** than actual number of entries/fills! In my case there were **10** points for **17** entries. I checked that, those specific bin contents were actually the sum of two entries! This is very weird!!! I don’t know what’s going on! Please advise.

Check out the figure below where I plot using both methods that illustrates the issue:

Blue triangle plot is produced by filling **Time** in the x-axis and bar profile is when I fill **Time + dt/2.**