[MENTION=6866167]Thurmas[/MENTION] I can see where you are coming from. When you have a concept in mind, you want to be able to realize that concept. But I think part of the problem of 5e and the archetypes is that it can give you an impression that every class should be good at every potential niche. This is especially true if you include all the options from Unearthed Arcana, SCAG, and XGtE. But I don't necessarily agree with that viewpoint. When I design material for my table, I pay very closely to balance within the class and archetype first, and then try to balance between classes. It is much harder to eyeball balance between classes, since each class is attempting something different. From my perspective, archetype add a flavor or spin to the class, but keep the character still grounded into its general niche. At the end of the day, a warlock is really, to borrow a term from 4e, a striker class. They have a few tricks that can keep them relevant in a number of situations, but they are always going to have a glass jaw. They can do well for a round or two going toe-to-toe with an enemy, but they aren't going to outlast the fighter or ranger. Now with some multiclassing tricks or through feats, they may be able to come close. But no patron or pact by itself should allow a warlock to be a proficient front-line melee character.
And this is the trap of the Hexblade. It makes the assumption that a warlock should be able to be a frontline melee character without feat investment, multiclass choice cost, and also be an expert in the use of Hex. It goes far and away from what any of the other patrons establishes for their characters. Each character gets their spin as a pet warlock, spell warlock, or weapon warlock, but none become experts or masters of those things. To do so, those warlocks still need invocations, feats, and particular spells to achieve their concept. They need time. But the Hexblade allows you to jump right in to its spin before pact is even considered.
I don't know. I kinda feel like I'm just rambling now, so I'm gonna stop.