I have a rather big expression which contains many sums of form q x + w y. I know that y is much smaller than x and i want to omit it where possible, namely if both q and w are integers. So i apply the replacement rule

q_?IntegerQ x + w_?IntegerQ y -> q x

However this rule obviously doesn't work if either q or w (or both) is equal to unity because FullForm of q x contains Times and one of x doesn't. So to get replacement done i need to use the ugly set of rules

{ q_?IntegerQ x + w_?IntegerQ y -> q x , x + w_?IntegerQ y -> x , ... }

and in place of dots there are two more rules with w and both w and q omitted.

So the question is whether it is possible (in general) to somehow simplify this ugly set to anything more simple.

  • $\begingroup$ BlankNullSequence or {q___?IntegerQ x + w___?IntegerQ y -> q x} $\endgroup$ – Timothy Wofford Sep 26 '13 at 11:50
  • 3
    $\begingroup$ Do you want 0s to match? What about q_. x + w_. y /; IntegerQ /@ And[q, w]? $\endgroup$ – Rojo Sep 26 '13 at 12:03
  • $\begingroup$ (Better And @@ IntegerQ /@ {q, w}, otherwise q=w=True would match) $\endgroup$ – Rojo Sep 26 '13 at 12:17

You can take advantage of the OneIdentity attribute of Times as it affects pattern matching(1),(2),(3) by making q and w Optional. (As Rojo shows in a comment above.) Condition is then used to check the q and w matches.

{3 x + 7 y, 2 x + E y, x + 5 y} /.
  q_. x + w_. y /; IntegerQ[q] && IntegerQ[w] :> q x
{3 x, 2 x + E y, x}

Note that I used RuleDelayed rather than Rule to localize q.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.