Adding Hysteresis to ADCMP380 Comparator

[ElektronikLabor]

Hi folks,
while designing a small circuit I came across to a beginner problem that I find challenging:

I found a low power comparator ADCMP380 that I want to use in a coin cell powered circuit (CR2032) that should last at least 5y.
That comparator has it's own reference voltage but unfortunately the non-inverting input is hardwired with it and there has no access to it; so I've got only the inverting input to work with.
On the inverting input I've got a voltage that is slowly rising (RC). When the threshold is reached the capacitor discharges until a lower limit and the cycle begins again (duration ~5min).
For that I need a hysteresis.

How can I add a hysteresis to the input?

According to the datasheet one need connect a resistor (R2) between the output and the lower end of the input voltage divider (see attached image).
In my opinion if you do that you get an oscillator: when the input becomes low when you reach the threshold, the voltage on the inverting input drops which switches off the output  .

Can you explain to me how that should result in a hysteresis?
Or is there a mistake in the datasheet?

Thanks

[xvr]

Output transistor will invert output signal, so negative input of comparator is effectively a positive one.

[ElektronikLabor]

Output transistor will invert output signal, so negative input of comparator is effectively a positive one.

Hi, thanks for quick reply
I still don't understand; although it's just a simple circuit I still can't figure out how that supposed to be a hysterersis.

[xvr]

When input voltage below threshold comparator output will be high and opened output transistor. Transistor in turn short output to ground.
So low input produce low output. And vice versa. So feedback to this input will be positive.

[PGPG]

I still can't figure out how that supposed to be a hysterersis.

After first xvr answer I supposed subject is closed.
My be you don't know that hysteresis need not be done at not used by you input but can be done at input you are using.
Just if when your inputs going up crosses the switching level and as a consequence signal seen at komparator input is (thanks to positive feedback) rapidly shifted more up than to reverse back the komparator state you have to go with your input more negative that the level at which previous state jump took place. So you get hysteresis.