Triangle region: dy dx to dx dy
Change the order of ∫₀¹ ∫ₓ¹ xy² dy dx and evaluate it.
- Region as a compound inequality:
- New limits:
- Inner integral:
- Outer integral:
Answer: , the same as the original order.

This change order of integration calculator swaps dy dx and dx dy in an iterated integral, rewrites the limits of integration, splits the region where a boundary changes and checks that both orders give the same value.
Results
Your region, sliced both ways
Switch between the original strips and the new ones. Each coloured band of the new order is one integral, and the dashed lines mark where the region has to be split.
Interactive diagram
Region of integration
Step-by-step solution
—Step by step
To use the change order of integration calculator, follow these 6 steps. The preview, the new integral and the region sketch update as you type.
Choose the order to change
dy dx → dx dy, dx dy → dy dx, or a triple integral with the two outer variables swapped.
Enter the integrand
For example exp(x^2) or x*y. Type 1 to work with the area of the region.
Enter the outer limits
Numbers only, such as 0 to pi.
Enter the inner curves
Curves in the outer variable, such as x^2 to x or 0 to sin(x).
Change the order
Press Change the order or Enter. Results refresh on their own while you type.
Read the new integral and the check
Compare the original and new strips on the sketch, then confirm both orders give the same value.
Definition
Changing the order of integration means rewriting an iterated integral so the inner and outer variables trade places, with new limits of integration that describe the same region. An integral in dy dx order becomes an integral in dx dy order, and the value stays the same.
A double integral over a region R adds up f(x, y) dA, where the area element dA equals dy dx or dx dy. Fubini's theorem turns that double integral into an iterated integral: one definite integral inside another. For a continuous function f on a rectangle R = [a, b] × [c, d], both orders have constant limits:
An iterated integral has 3 parts that change in a swap:
On a rectangle, the swap only exchanges the differentials and the limit pairs. On any other region, the inner limits are curves, and the new limits come from the graph of the region. The order of variables changes; the region, the integrand and the answer do not. The area of R (integrand 1), the volume of the solid under a positive function f, and the average value favg = (1/Area)·∬R f dA come out equal in both orders.
A vertical strip has fixed x, so its ends are y-values. A horizontal strip has fixed y, so its ends are x-values. Switch the order to see the limits change while the region stays put.
Reasons
There are 2 reasons to change the order of integration: the inner integral has no elementary antiderivative in the first order, or the limits are simpler in the other order. A third use is checking work, since both orders give the same number.
An integral with no elementary antiderivative in the first order becomes solvable after the swap when the new inner integral treats the hard function as a constant. ex², sin(x²), cos(x²), e−y² and √(1 + y³) have no antiderivative made of elementary functions, and no table of integrals lists one.
Integrate with respect to the other variable first, and the hard function rides along as a constant. The inner integral then multiplies it by a polynomial, and that polynomial is the derivative of the expression inside. A substitution with t and dt finishes the outer integral.
The new inner indefinite integral still uses the 5 single-variable rules: the constant multiple property of integrals, the sum rule for integrals, the difference rule for integrals, the substitution rule and integration by parts.
| Hard order | After the swap | Value |
|---|---|---|
Simpler limits in the other order mean fewer integrals to evaluate. A region with 2 different top boundaries needs 2 integrals in dy dx order, but only 1 integral in dx dy order when its left and right boundaries stay the same curves from bottom to top.
The tent triangle with corners (0, 0), (1, 1) and (2, 0) shows the difference. Its top edge changes from y = x to y = 2 − x at x = 1, so vertical strips need 2 integrals. Horizontal strips always run from x = y to x = 2 − y, so 1 integral covers the whole region:
Pick the order whose strips cross the region between the same 2 curves from start to finish.
Red: a vertical strip, which ends on y = x or y = 2 − x depending on the side. Teal: a horizontal strip, which always ends on x = y and x = 2 − y.
Method
To change the order of integration, write the region as inequalities, sketch it, read the new constant outer limits off the sketch, then solve each boundary curve for the other variable. The 4 steps below change the order of ∫₀² ∫x²4 f(x, y) dy dx.
To write the region as inequalities, copy each limit pair onto its own variable: the inner limits bound the inner variable, and the outer limits bound the outer variable.
Some pairs merge into one compound inequality. The triangle ∫₀¹ ∫ₓ¹ f dy dx gives 0 ≤ x ≤ y ≤ 1, and the new description 0 ≤ x ≤ y with 0 ≤ y ≤ 1 reads straight off it.
To sketch the region, draw each boundary as a curve and shade the side the inequalities allow. This region sits above the parabola y = x², below the line y = 4 and to the right of the y-axis x = 0. Mark the corner points (0, 0), (2, 4) and (0, 4); corner points supply the new outer limits and any split points.
To find the new constant outer limits, project the region onto the axis of the new outer variable and read the lowest and highest values. The region spans y = 0 to y = 4, so the new outer integral runs from 0 to 4. Outer limits are always constants: a variable in an outer limit turns the answer into a function instead of a number.
To solve the boundary curves for the other variable, draw a strip for the new inner variable and name the curve where it enters and the curve where it leaves. A horizontal strip at height y enters at x = 0 and leaves at the parabola. Solving y = x² for x gives x = √y, the positive root, since x ≥ 0 in this region.
With f = 1, both orders give the area 16/3 ≈ 5.3333 square units: 5.3333 cm² for lengths in centimeters, or about 0.8267 in².
Region types
A Type I region lies between 2 functions of x over an x-interval, and a Type II region lies between 2 functions of y over a y-interval. Type I matches dy dx order with vertical strips; Type II matches dx dy order with horizontal strips. Changing the order of integration describes the same region as the other type.
Many regions are both types at once, including rectangles, triangles, disks and the lens between y = x² and y = √x. A region that is one type as a single piece can need several pieces as the other type, and those pieces become separate integrals after the swap.
| Type I | Type II | |
|---|---|---|
| Description | ||
| Strips | vertical | horizontal |
| Order | dy dx (integrate y first) | dx dy (integrate x first) |
| Iterated integral | ||
| Lens example |
Pieces
A region must be split into pieces when the new strips enter or leave the region through different curves in different parts of the range. Each change of boundary formula starts a new integral, and the pieces add up to the original value.
The region under y = eˣ for 0 ≤ x ≤ 1 shows the rule. Below y = 1, a horizontal strip runs from the y-axis x = 0 to the line x = 1. Above y = 1, the strip starts on the curve x = ln y. The new integral has 2 parts:
Split points come from the corners of the region, where one boundary curve meets another. Here the corners sit at y = 1, where eˣ meets the y-axis at (0, 1), and at y = e, the point (1, e). The calculator finds these points and writes one integral per piece; the valley curve y = x² − 2x + 2 on [0, 3] splits into 3.
Green: piece 1, a unit square. Teal: piece 2, from ln y to 1. The strip's left end jumps from the y-axis to the curve at y = 1.
Triple integrals
A triple integral has 6 orders of integration, since 3 variables arrange in 3! = 3 × 2 × 1 = 6 ways: dz dy dx, dz dx dy, dy dz dx, dy dx dz, dx dy dz and dx dz dy. Each order needs its own limits, and all 6 give the same value for a continuous integrand.
The method repeats the 2D steps on a projection. Project the solid onto a coordinate plane, change the order of the 2 outer variables in that plane, then keep or rebuild the inner limits. The tetrahedron in the first octant under x + y + z = 1 has volume 1/6 in every order.
The calculator's triple mode keeps the inner z-limits and swaps the outer pair, dz dy dx → dz dx dy, the most common textbook change. To move z out of the inner position, sketch the projection onto the xz- or yz-plane first. Cylindrical coordinates (r, θ, z) and spherical coordinates (ρ, θ, φ) follow the same count: 6 orders each.
| Order | Innermost | Middle | Outermost |
|---|---|---|---|
| dz dy dx | |||
| dz dx dy | |||
| dy dz dx | |||
| dy dx dz | |||
| dx dy dz | |||
| dx dz dy |
Worked examples
Here are 4 worked examples of changing the order of integration, in the style of the Math Insight examples: a triangle, a region between 2 curves, an ex² integral that only works after the swap, and a region that needs 2 pieces. Red strips show the old order; teal strips show the new one.
Change the order of ∫₀¹ ∫ₓ¹ xy² dy dx and evaluate it.
Answer: , the same as the original order.
Change the order of ∫₀¹ ∫x²√x xy dy dx.
Answer: . The lower curve in one order becomes the upper limit in the other.
Evaluate ∫₀¹ ∫y1 ex² dx dy. The inner integral has no elementary antiderivative, so change the order.
Answer:
Change the order of ∫₀¹ ∫₀eˣ dy dx, the area under y = eˣ.
Answer: , equal to .
Watch out
There are 8 common mistakes when reversing the order of integration. Each card pairs the mistake with its fix.
Swapping the integral signs only
∫ₐᵇ ∫g(x)h(x) f dy dx does not become ∫g(x)h(x) ∫ₐᵇ f dx dy. Rebuild every limit from the region.
Variables in the outer limits
Outer limits are constants. A variable there leaves a function instead of a number.
Skipping the sketch
The new limits come from the graph of the region, not from the old symbols. Mark the corner points first.
Missing a split
Start a new integral wherever the new strip changes boundary curve: under y = eˣ, at y = 1.
Wrong root when inverting
y = x² gives x = √y where x ≥ 0 and x = −√y where x ≤ 0. Check the sign against the sketch.
Left and right limits reversed
A decreasing curve such as y = 2 − x becomes x = 2 − y on the right side, not the left.
Differentials out of step with the limits
The inner limits belong to the first differential. ∫∫ … dx dy means the inner limits are x-values.
Swapping a function that is not integrable
(x² − y²)/(x² + y²)² on [0, 1] × [0, 1] gives π/4 in one order and −π/4 in the other. Check that ∬|f| dA is finite.
Questions
No, changing the order of integration does not change the answer when f is continuous on a bounded region, or when the integral of |f| over the region is finite. Fubini's theorem guarantees the same value in every order. The rule fails for functions that are not absolutely integrable: (x² − y²)/(x² + y²)² over the unit square [0, 1] × [0, 1] gives π/4 in dy dx order and −π/4 in dx dy order.
The easier order is the order whose inner integral has an elementary antiderivative and whose region needs the fewest pieces. Check the integrand first: e^(x²) integrated with respect to x has no elementary antiderivative, so put dy inside. Then sketch the region and count the boundary curves each strip crosses. General solvers such as Wolfram Alpha, Symbolab and Mathway evaluate the order you type; this calculator rewrites the order and evaluates both.
Fubini's theorem states that a double integral of a continuous function f over a rectangle R = [a, b] × [c, d] equals either iterated integral: integrating f dy dx over y from c to d and x from a to b gives the same number as integrating f dx dy over x from a to b and y from c to d. Guido Fubini published the general version in 1907. The theorem extends to any region that is both Type I and Type II, and to any function with a finite integral of |f|. Tonelli's theorem covers nonnegative functions.
The new integral needs 2 parts when the new strips start or end on different curves in different parts of the region. The region under y = eˣ for 0 ≤ x ≤ 1 needs 2 parts in dx dy order: below y = 1, a horizontal strip runs from x = 0 to x = 1; above y = 1, the strip starts on the curve x = ln y. The 2 parts add up to 1 + (e − 2) = e − 1 ≈ 1.7183.
A triple integral has 6 orders of integration: dz dy dx, dz dx dy, dy dz dx, dy dx dz, dx dy dz and dx dz dy. 3 variables arrange in 3! = 3 × 2 × 1 = 6 ways, and each order needs its own limits. The tetrahedron x + y + z ≤ 1 in the first octant has volume 1/6 in all 6 orders.
This change order of integration calculator swaps dy dx and dx dy, rewrites the limits of integration from the region, splits it into pieces where a boundary changes, and checks that both orders give the same value.