Minioncookies is correct in saying that the long list of side-effects is because some people may experience any one of those effects. And since the FDA requires drug manufacturers to list every side-effect that was ever reported, the list can be quite exhaustive. But that doesn't mean that every person taking those drugs will get every side-effect listed (nor even getting close to half of them).
Another thing that causes the seemingly paradox of how an antidepressant may lead to suicidal tendencies is due to the mechanism of action for that drug (it has been quite a while since I studied the pharmacological effects of antidepressants, so I may not have this entirely correct, but....). Some theories as to what causes depression is that the neurotransmitter serotonin is imbalanced in the brain. The brain has this feedback mechanism whereby if there is too much serotonin being passed between cells, there is a serotonin reuptake receptor that will essentially "soak up" (or reuptake) the excess serotonin. But sometimes (it is believed) that this mechanism is malfunctioning and there are too many reuptake receptors causing levels of serotonin to be low. So they have developed reuptake inhibitors in order to increase the amount of serotonin in the gap (or synapse) between neurons (or brain cells). But remember that I said that the reuptake receptors were a part of a feedback mechanism for removing excess serotonin? In addition to just removing the excess serotonin, it is also thought that it goes on to inhibit production of serotonin by those cells that reuptake it. If that is also the case, then if a person is given a serotonin reuptake inhibitor, it will decrease the reuptake of serotonin but it doesn't effect the production, so that when the serotonin in the synapse is used to transmit its signal to the next neuron, the synapse will once again become deficient of serotonin levels until the feedback loop is broken and serotonin production goes back to normal. And it is because of that temporary low levels of serotonin that may increase the risk of suicidal tendencies.
Finally, neuroscience is one of the fields that is very difficult to study considering the vital importance of that organ, the blood-brain barrier, difficulty in measuring neurotransmitter levels, etc. So even though certain drugs seem to help in the management of depression, there is still a lot of unknowns about the exact mechanism of how they work. So when you say that doctors are supposed to know what a drug will do for a person, their knowledge is only to the point that neuroscientists and pharmacologists know what is actually happening. If the scientists working on these type of medications don't fully know what is going on, how is a doctor supposed to know?!
But just because we don't fully understand the mechanism of a drug doesn't mean that they shouldn't be used either. That is why drugs have to go through a gamut of clinical trials before the FDA will approve their use. Phase 1 clinical trials determines the toxicity of a drug, phase 2 tests whether a drug actually works the way that it is intended to work (as well as to gauge the risk-to-benefit factor), and phase 3 is to see how effective the drug is compared with current treatments. So if an antidepressant is shown to help 95% of those that take it while increasing the risk of suicidal tendencies in only 1%, the risk-to-benefit ratio would suggest that this drug is an effective treatment even though it does seem to have the opposite effect in a slight minority of those taking it. With cancer chemotherapy, this risk-to-benefit ration is sometimes only to look at the longevity of a person rather than its lethality; if a drug is found to increase a person's length of life by a few years even though it is known that the drug will wind up killing the person later on, that increase in longevity may be enough to tip the scale in having doctors prescribe that esp. if it is the only medication currently available to treat that particular form of cancer.