How is it possible that row cultivation was not widely practiced in Europe before the 18th century?

If you want to grow anything, the most logical and simple thing would be to grow it in rows, right?

Why? The most logical and simple thing is to imitate nature, which doesn't use rows.

What is important is to have the plants suitably distributed - not too close, and not too far apart. If you have a cereal crop where the plants like to be a few centimeters apart, say about 5cm, then if they are evenly distributed, you will have about 400 plants per square metre. How do you fit those same plants into rows 25cm apart without crowding them too close together in the rows? In practice, you can crowd them a bit along the row, and the extra space between the rows makes up for it. But this isn't obvious. For plants that can grow close together, row cultivation looks like it wastes a lot of space. Does it bring benefits? Row cultivation allows easier weeding (e.g., using hoes, or by machine). Row cultivation makes mechanised spreading of fertiliser, application of herbicides, insecticides, etc. easier. If weeds are not likely to be a great problem, and the farming is unmechanised, there is less benefit from row cultivation.

The other issue is planting the seeds. For crops that need a lot of space per plant, it's easy to plant seeds (or seedlings) individually, or in groups of 2-3 per mound (to be thinned to 1 later) or similar. This will typically result in rows of plants. See, for example, this turnip field, c. 1400:

https://it.wikipedia.org/wiki/File:Ciclo_dei_mesi,_settembre.jpg

For plants that will be sown at higher densities, individual planting adds a lot of labour. It is sometimes done for cereals. For example, rice seedlings are traditionally planted by hand. Sometimes, they are planted in fairly neat rows, often in irregular rows, sometimes approximately a uniform grid, and sometimes in a random grid. For wheat, barley, oats, rye, the traditional method of sowing was broadcasting, spreading grain randomly (but as uniformly as possible) by hand. If you try to plant a wheat field by hand, individual seed by individual seed, the merits of broadcasting very quickly become clear. On the other hand, broadcasting has disadvantages. Randomly spreading seeds does not produce a uniform distribution of plants - many plants will be clustered together, and will be closer to each other than ideal. There will be gaps. These will occur just as a result of randomness. Gaps and clusters will either reduce yields or increase the amount of seed used to get good coverage; neither of these is desirable. More uniform distribution of plants can be achieved by controlled planting in rows. For wheat etc., this only became a standard pattern of sowing with mechanisation, with the introduction of the seed drill. The seed drill comes with costs - one needs to buy and maintain the machine, and have animal or mechanical traction available to pull it. If one's crops are spread over small dispersed fields (a common and useful measure to try to avoid total crop failure, at the cost of lower average yields), this adds further obstacles to mechanisation. Even in parts of England using modern farming methods, broadcasting of cereals remained common into the 19th century.

In some places, there has been a return to broadcasting. For example, in large-scale mechanised wet rice farming, the rice is sometimes sown by being broadcast from aircraft. For rice, aerial broadcasting, mechanical drill-sowing, and hand-planting of seedlings are all used in different places. Cost of labour compared to cost of buying, maintaining, and operating machinery (seed drills, tractors, aircraft), and the amount of land and expected yield per hectare combine to determine which is the best method, economically speaking.