Here are two additional facts worth noting about this technique of "reverse engineering" a factorization from the zeros:
First, it will always be possible to clear fractions in this situation (namely, where we are trying to factor a quadratic with integer coefficients and rational zeros). We know this because we know it can be factored over the integers, and that the linear factors must be multiples of the rational factors you have formed, so there are linear factors with integer coefficients that we obtain by multiplying -- namely, by factors of a.
Second, we can get directly to the proper factoring over the integers with a slightly different approach. We have found the zeros 1/2 and -2/3; rather than using x - 1/2 and x + 2/3 as the factors, we can instead directly write 2x - 1 and 3x + 2, since if x - 1/2 = 0, then 2x - 1 = 0 by multiplying by 2. That is, for a rational zero m/n, we can write factor nx - m, which equals 0 when x = m/n. Therefore, we know that (2x - 1)(3x + 2) is a quadratic polynomial with the same zeros as 6x^2 + x - 2; and any quadratic polynomial with the same zeros and the same "a" is, in fact, the same polynomial. So this method will always work.
Finally, we should keep in mind that what is commonly just called "factoring" in elementary texts is actually "factoring over the integers", so that integer coefficients are expected; 6(x - 1/2)(x + 2/3) typically will not be accepted, even though in other contexts it is indeed a valid factorization. We don't know for sure the context of the OP.