updated description:

You are not restricted to plain SQL. You may use PLPGSQL dialect.

That is a good remark. On the other hand, you are never restricted to plain SQL in a kata.

Thanks a lot. Converting rental_date to date type corrected the situation. I overlooked that rental_date was of type time_stamp. Probably because of this, the comparison with the date type was incorrect. I'll keep it on mind.

Python fork

• Fixed test framework setup
• Added assertion message
• Description fix

Very cool

The error in the path description:
For x = 1 and y = 1, the output should be 1.
Path:(0,0) --> (0,1) --> (1,1), 1 turn at (0,1)
Coordinates of the second point in the path are wrong (0,1). The correct coordinates is (1,0). The correct path is: Path:(0,0) --> (1,0) --> (1,1), 1 turn at (1,0).
The same error in the second example.

error in the test is impossible because before approval kata is being solved by the smartest 2-1 Dan guys who detect all possible errors if they are there

I completely agree. Debugging was torture. Displaying the query result is required. There is no normal diagnostic information in the task.

Bad non-deterministic tests. The same solution failed the test with an incomprehensible error. At the next test of the same solution, suddenly the solution was passed. The result of the query is not displayed, the difference between the solution and the correct solution is not displayed in a tabular form with error highlighting. As a result, there is no normal diagnostic information for debugging.

The description isn't complete. The data schema of the "prospect" table is not fully specified - it has an "id" column that is not specified in the description.