To find the inverse of a function from a table, swap every (x, y) pair into (y, x) and then check whether the result still passes the definition of a function. If any y-value in the original table repeats with different x-values, the inverse is not a function — and no amount of rearranging will change that. The method takes about a minute per table: copy the pairs, reverse them, re-sort by the new input, and verify one-to-one behavior. The tables below show the full procedure on a one-to-one table that passes and a many-to-one table that fails.
Why the Swap Works and Where It Breaks
An inverse relation undoes a function by sending each output back to the input that produced it. Write the original rule as y = f(x). The inverse, written f⁻¹, satisfies f⁻¹(y) = x for every pair the table lists. Reversing the columns is not a trick — it is the definition applied column by column. That's why the first move in every table problem is mechanical: copy and flip.
The catch sits in the word "function." A function assigns exactly one output to each input. After swapping, the old outputs become the new inputs. If the original table gave the same y to two different x-values, the reversed table hands that new input two outputs — and the inverse is not a function. The classic example is f(x) = x², where f(2) = 4 and f(−2) = 4. Reverse those pairs and 4 maps to both 2 and −2.
Graphically, this is the horizontal line test. A function has an inverse that is also a function only when no horizontal line crosses its graph more than once. For a table, you run the same test by scanning the y-column for duplicates. One repeated value with different partners kills the inverse outright; repeated y-values always come from distinct x-values, so any duplicate in that column is decisive.
Terminology matters here. The inverse relation always exists — you can always swap pairs and plot them. What can fail is the inverse function. Textbooks such as the OpenStax Algebra and Trigonometry 2e (Section 3.7) keep the two terms separate for exactly this reason, and teachers grade on it. If a prompt asks for "the inverse" without qualification, state whether the result is a function and why.
Read the Table's Domain and Range Before Touching the Pairs
The x-column of the original table is the domain of f, and the y-column is its range. After inversion, those roles trade places: the inverse's domain is the original range, and the inverse's range is the original domain. Skipping this step causes errors later when you state restrictions, so do it first with the raw numbers in front of you.
Take this table:
| x | 1 | 2 | 3 | 4 |
|---|---|---|---|---|
| f(x) | 3 | 5 | 9 | 12 |
The domain of f is {1, 2, 3, 4}. The range is {3, 5, 9, 12}. Those four values become the domain of f⁻¹, and {1, 2, 3, 4} becomes its range. Write them down before swapping so the final answer includes them.
Set notation is the safe format: domain {3, 5, 9, 12}, range {1, 2, 3, 4}. Ordered lists of endpoints invite mistakes when the values aren't consecutive, and interval notation doesn't fit discrete table data at all. Some curricula accept a simple list, so check which format your course uses.
Domain and range also expose repeated outputs early. If the y-column shows 3 twice, you already know the inverse won't be a function, and you can say so before doing any swapping. This pre-check saves time on exams where the question is "is the inverse a function?" rather than "find it."
Step 1: Copy Every Ordered Pair Exactly
Write the table as a list of ordered pairs before changing anything. Accuracy at this stage prevents every downstream error, and it takes seconds. The pairs from the table above are (1, 3), (2, 5), (3, 9), and (4, 12).
Read left to right, top to bottom, and say each pair aloud if you're working by hand. Verbalizing catches transposed digits — writing (3, 1) when the table says (1, 3) is the single most common mistake in this procedure. Most published worksheets print the table horizontally, with x on top and f(x) below, which makes row-by-row pairing easy to scramble.
Check the pair count against the table. Four columns should produce four pairs; if you have three, you dropped one. Some textbook tables include a "y" row with one blank cell waiting to be filled. Treat blanks as data to compute, not data to ignore. And if the table arrives with columns already labeled y and x, you're looking at the inverse — say so rather than inverting twice.
Step 2: Reverse the Order in Each Pair
The swap is literal: (a, b) becomes (b, a). Applied to the running example, (1, 3) becomes (3, 1), (2, 5) becomes (5, 2), (3, 9) becomes (9, 3), and (4, 12) becomes (12, 4). The new table reads x: 3, 5, 9, 12 and f⁻¹(x): 1, 2, 3, 4.
Keep the pairs aligned while you work. A two-column table with the new x on the left and the new f⁻¹(x) on the right is the standard format, and it's the one most teachers expect on paper. Don't skip straight to a scatter of numbers; the pairing must stay visible.
At this point the inverse is a relation, not yet a verified function. The next two steps settle that question. Don't write "f⁻¹(x) =" until the duplicate check is done — that notation claims function status, and if the check fails you'll have to take it back.
Step 3: Sort the New Inputs and Check for Duplicates
Most textbook tables list x-values in increasing order, so the reversed table often runs out of order. Re-sort the swapped pairs by their new input, smallest to largest, then scan for duplicates. Sorting is bookkeeping, not mathematics, but it's how you catch a broken inverse.
Here's the original table and its inversion side by side, including a deliberately unsorted case:
| Original x | Original f(x) | Swapped pair | Sorted new x | f⁻¹(x) |
|---|---|---|---|---|
| 1 | 3 | (3, 1) | 3 | 1 |
| 2 | 5 | (5, 2) | 5 | 2 |
| 3 | 9 | (9, 3) | 9 | 3 |
| 4 | 12 | (12, 4) | 12 | 4 |
Now the failing case, where f(x) = x² on the inputs −2, −1, 0, 1, 2:
| Original x | Original f(x) | Swapped pair | Duplicate output? |
|---|---|---|---|
| −2 | 4 | (4, −2) | Yes |
| −1 | 1 | (1, −1) | No |
| 0 | 0 | (0, 0) | No |
| 1 | 1 | (1, 1) | Yes |
| 2 | 4 | (4, 2) | Yes |
Two new inputs, 1 and 4, each carry two outputs. The inverse relation is {(4, −2), (1, −1), (0, 0), (1, 1), (4, 2)}, and it is not a function. No sorting fixes this; the duplicates are structural.
A time computation worth internalizing: on a five-row table, sorting and scanning takes under 60 seconds, versus roughly 15 minutes of rework when a duplicate slips through into a graded answer. The check is cheap relative to the failure it prevents.
Step 4: Confirm It's a Function and Write the Notation
If every swapped input appears exactly once, the inverse is a function. Write it as f⁻¹(x) with the new pairs beneath, and state the domain and range you recorded earlier. For the running example: f⁻¹ = {(3, 1), (5, 2), (9, 3), (12, 4)}, domain {3, 5, 9, 12}, range {1, 2, 3, 4}.
If duplicates appear, report the inverse relation and say plainly that f⁻¹ is not a function. Some curricula accept "the inverse is not a function" as a complete answer; others want the relation listed anyway. Do both — it costs a line and removes ambiguity.
The notation f⁻¹(x) does not mean 1/f(x). That misunderstanding is common enough that the College Board's AP Precalculus materials flag it directly. f⁻¹ is the undoing map; 1/f is a reciprocal. They are unrelated objects, and mixing them up loses points even when the table work is correct.
Verify With Composition: f(f⁻¹(x)) = x
Verification is arithmetic: feed a swapped pair back through the original table and confirm you land on the starting value. Take (3, 1) from the inverse. The original table maps 1 to 3, so f(f⁻¹(3)) = f(1) = 3. One check per pair is enough to catch a copy error.
The general identity is f(f⁻¹(x)) = x for every x in the inverse's domain, and f⁻¹(f(x)) = x for every x in the original domain. For table data you don't need the formulas — you look values up in the other table. Go through all four pairs of the running example and each round trip should return the input.
If any round trip fails, the error is almost always in Step 1 or Step 3, not in the concept. Re-copy the original pairs and re-sort the swapped ones. Composition failures in table problems are bookkeeping failures roughly every time.
When the Inverse Doesn't Exist as a Function, and How to Repair It
A many-to-one table has no inverse function, but it may have an inverse function once you restrict the domain. The quadratic example above becomes invertible if you keep only x ≥ 0, giving the pairs (0, 0), (1, 1), (2, 4). Now each output is unique, and the inverse {(0, 0), (1, 1), (4, 2)} is a function.
Restriction is the standard repair, and it's what high school and early college courses expect when a bare table fails the horizontal line test. Read the problem's stated domain first; many tables already restrict inputs, which is why some quadratic tables invert cleanly and others don't.
Don't invent a restriction that contradicts the given data. If the table includes both (2, 4) and (−2, 4), dropping one row changes the function, and that must be justified by the context — a physics problem where negative time is meaningless, for example. State the restriction explicitly in your answer.
Common Mistakes With Table Inversions
The errors below account for most lost points on this topic, and each has a specific fix.
| Mistake | What it looks like | Fix |
|---|---|---|
| Transposing digits during the copy | (3, 1) written for (1, 3) | Read pairs aloud; check count against columns |
| Skipping the duplicate check | Writing f⁻¹(x) = when a y-value repeats | Scan the y-column before swapping |
| Reading f⁻¹ as a reciprocal | Treating f⁻¹(2) as 1/f(2) | f⁻¹ undoes f; test with composition |
| Forgetting to swap domain and range | Listing the original domain as the inverse's | Record both before Step 1 |
| Inverting a table that is already an inverse | Swapping back to the original | Check column headers first |
| Averaging repeated outputs | Writing (1, 0) from (1, −1) and (1, 1) | Report both pairs; the relation isn't a function |
The averaging error deserves a note because it feels mathematically reasonable and is wrong. An inverse relation keeps every pair; it never collapses duplicates into a middle value. If (1, −1) and (1, 1) both appear, the inverse contains both, and the inverse is not a function.
A subtler trap: tables where the y-column has repeated values but the question only asks for the inverse relation. Read carefully. "Find the inverse" without "function" may accept a relation, and marking it "not a function" as a refusal to answer loses credit.
Checking the Inverse Against Its Graph and Worked Answers
Plotting both tables on the same axes makes the swap visible: every point of the inverse is the original point reflected across the line y = x. The pair (2, 5) reflects to (5, 2), and the line y = x passes through the midpoint of the segment joining them. This reflection is a reliable visual check when you have graph paper or a graphing tool.
The reflection also explains the duplicate problem geometrically. A function whose graph fails the horizontal line test has two points at the same height; reflecting them produces two points at the same horizontal position, which is the vertical line test failing on the inverse. Table duplicates and graph failures are the same fact in two formats.
If a key or answer sheet gives f⁻¹ as a formula rather than a table, test the formula against your swapped pairs by substitution. A formula that reproduces all four swapped pairs is consistent with the table, though a four-point table doesn't prove the formula for every input. For exam purposes, matching the given pairs is the standard of evidence.
FAQ
What if the table has the same y-value twice with different x-values?
The inverse relation exists but is not a function, because that shared output becomes an input with two outputs after the swap. List both reversed pairs and state that f⁻¹ fails the definition of a function, or restrict the domain if the problem allows it. No rearrangement of the table changes this.
Does f⁻¹(x) mean 1 divided by f(x)?
No — f⁻¹ is the inverse function, which undoes f, while 1/f(x) is the reciprocal. For f(x) = x + 1, the inverse is f⁻¹(x) = x − 1, but the reciprocal is 1/(x + 1). The two produce different values for almost every input, and AP Precalculus materials single out this confusion as a recurring error.
How do I find the inverse if the table lists x-values out of order?
Sort the original pairs by x before swapping, then sort the swapped pairs by the new input. Order doesn't affect the mathematics, but sorted columns make the duplicate check mechanical and match the format most answer keys use. An unsorted table is still invertible; it just takes one extra pass.
Can a table's inverse be a function if the original table has only three rows?
Yes. Row count is irrelevant; uniqueness of the y-values is what matters. A three-row table like (1, 4), (2, 8), (3, 12) inverts to a function because no output repeats. Conversely, a ten-row table with one repeated output fails regardless of length.
What's the quickest way to check my swapped table is right?
Run one composition: pick a swapped pair (b, a), look up a in the original table, and confirm it returns b. If two or three random pairs round-trip correctly, the swap is almost certainly clean. A failed round trip means a copy or sort error, not a conceptual one — go back to Step 1.
Is the inverse of a linear table always a function?
Not automatically. A linear function with a nonzero slope is one-to-one, so its table inverts to a function. But a table of a horizontal line, such as (1, 5), (2, 5), (3, 5), has a repeated output and no inverse function. Check the y-column rather than assuming from the pattern.
Why does reflecting across y = x matter for tables?
Each swapped pair is the original point mirrored over y = x, so the inverse table and the reflected graph contain the same points. If a plotted inverse doesn't look like the original graph flipped over that line, one of the tables has an error. It's the fastest visual audit available without recomputing anything.
Can I average two repeated outputs to make the inverse a function?
No. Averaging invents a pair that appears in neither the original table nor its inverse, so the result is mathematically wrong. The inverse relation keeps every pair, and the correct answer states that the relation fails the function test rather than fabricating a single midpoint value.