Some integrands have no direct standard formula but can be decomposed into a sum or difference of functions whose integrals are known, then integrated term by term. Typical decompositions: expanding a power such as (3x−1)2, splitting a single fraction like xx2−3x+1 into separate terms, and using trigonometric product-to-sum identities (e.g. 2cosAsinB=sin(A+B)−sin(A−B), cos3x=41(3cosx+cos3x)), power-reduction (1+cos2x=2cos2x, 1−cos2x=2sin2x), and sin2x+cos2x=1.
Partial fractions. When the integrand is a rational function q(x)p(x) with degp<degq (if not, first do polynomial long division), resolve it into a sum of simpler fractions — one term per distinct linear factor x−αA (and (x−α)2B for a repeated factor) — determine the constants by comparing coefficients or substituting convenient x-values, then integrate each simple fraction (each gives a logarithm or a power).
If the numerator's degree is ≥ the denominator's, you must divide first; only the proper-fraction remainder is expanded into partial fractions.