G7MRV:
So long as the switch contacts are clean, the energy dissipated by the actual switch will be so negligable as to be practically unmeasurable. With the switch close you have effectively just another wire.
For almost any practical purpose a close mechanical switch can be considered to use no energy.
A switch with dirty contacts however will high higher resistance and that would lead to some power being lost as heat, but its likely that your electronics would fail to register the switch and your code stop working properly long before the energy lost became significant
I highly recommend not answering questions if you don't know what you're talking about and are incapable of basic reasoning. I'm usually not this much of a dick, but your entire post contains literally nothing worth reading and is just wrong.
While it is true that an ideal switch will never dissipate power (since it will either have infinite resistance and pass no current, or 0 resistance and drop no voltage), that is an irrelevant fact. Given that OP is concerned about a battery powered circuit, the important metric to be concerned about is the delta of current consumption of the whole system when the switch state changes. And actuating the switch really does make a difference to that.
SPST switches used as inputs to a digital circuit require a pullup (or pulldown) resistor, and that will have current drawn through it when the switch is closed. Selecting the right resistor value becomes a balancing act between noise rejection (smaller is better) and lowering power consumption (higher is better).
So now to helping the OP:
The internal pullups in the Arduino Uno are between 40k-60k. Worst case, each switch will pull out 125 uA of current when closed. This might not seem like a lot, but considering that the processor itself can easily get down to single-digit microamps or less by using sleep modes it's quite significant if you're trying to conserve power. If you have several switches that are going to be closed for long periods of time, those few hundred extra microamps can blow your power budget right out of the water depending on how stingy you want to be.
There are other options. You can reduce the frequency of reading the switch. Human reaction time is extraordinarily slow by the standards of a microprocessor, even a low-end one like an Uno. For inputs that are part of a human interface, as long as you respond within 100 ms (an eternity in microprocessor-land) a human will perceive that as instantaneous. You can do something like only checking the switch every 5 ms and keeping the pullup resistor disconnected when you aren't reading the switch. That way even if the switch is held closed, it draws no power when you aren't reading it.
However, all this is academic so far. Have you calculated your power budget? What type of battery are you using? What is the voltage and capacity? Are you using any voltage regulators? What amount of lifetime are you trying to achieve? What other parts of your system are using power from the battery, and how much?
Until you have a proper power budget, you're just flailing around in the dark. If you don't know how much power is being using in what parts of the system, you end up being "penny wise, but pound foolish". If you do something stupid like dump 20 mA into a "power on" indication LED, there's not much point fussing about a few hundred microamps somewhere else. Always address the biggest problems first. If you don't know which component is your biggest problem, then that's a problem in and of itself.
should i just use a normal button instead of the switch ?
Switches and buttons are all fundamentally the same, just two pieces of metal that are pressed together (NO) or separated (NC) when the switch is actuated. The type of actuator (button or level or something else) doesn't really matter.