Texas has 254 counties. The next-highest state in the same dataset has 157, North Carolina has 100, and the median state has around 60, so any county-keyed structure in Texas is not a slightly larger version of the same problem — it is a different problem. A dropdown with 254 options is a usability decision rather than a form field, and a tax or jurisdiction table keyed on county carries more Texas rows than most states have addresses in a test fixture.
The postal spread is proportionate: 2,600 postal codes across 52 three-digit prefixes, and 1,469 distinct USPS place names — more distinct place names than most states have postal codes. A fixture set that draws from Houston, Dallas and Austin covers three prefixes out of 52 and misses the entire rural half of the state, where a single place name can cover a very large area.
Texas also owns one of the rule-breaking prefixes. Across the whole dataset only 6 of 926 three-digit prefixes appear in more than one state, and 739 is one: it spans Oklahoma and Texas. That is a real border case to have in a validation suite, because the usual assumption — a prefix belongs to exactly one state — holds for 920 of the other 926.
For a shipping or tax fixture the practical advice is to spread destinations by prefix rather than by city. Texas is wide enough that two addresses in the same state can sit in different zones for every carrier, and picking three metros gives a suite that agrees with itself and with nothing else.
The county count has a second-order effect worth planning for. Local sales tax in Texas is levied by cities, counties, transit authorities and special purpose districts, and their boundaries do not follow postal ones — which is why the state publishes an address-level rate lookup rather than a ZIP-level table. Any implementation that maps a five-digit ZIP to one rate is approximating in Texas, and the approximation is wrong most often exactly where the districts overlap. If your checkout claims to compute Texas tax from a ZIP, that claim is the thing to test.