Intuition What this page is for
The parent note gave you the rule. This page hunts down every situation a one-sided-limit problem can throw at you and works one example for each. Nothing here is new theory — it is a tour of cases . When you finish, no exam version should surprise you, because you will have seen its "shape" already.
How to read each example: first cover the answer and make a Forecast (a guess). Then walk the numbered steps — each says why we did it. Then Verify by plugging back or sanity-checking.
Every one-sided-limit question is really one of a handful of shapes . Here is the full list, and which example kills each.
Cell
Case class
What makes it tricky
Example
A
Sides agree (piecewise)
must check both rules
Ex 1
B
Sides disagree, both finite (jump)
equal-existence trap
Ex 2
C
Domain edge — one side undefined
only ONE side exists
Ex 3
D
Pole / infinite — sides go to ± ∞
sign of denominator
Ex 4
E
Floor / step , tricky non-integer twist
floor jumps only at integers
Ex 5
F
0/0 removable — algebra needed first
can't just plug in
Ex 6
G
Squeeze at a wiggle (sin x 1 type)
bounded-times-small
Ex 7
H
Word problem (price/tax jump)
translate story → piecewise
Ex 8
I
Exam twist — solve for an unknown constant
force sides to match
Ex 9
We build the vocabulary from zero first, then march through A→I.
Definition The three symbols we will lean on
x → a + means "x slides toward the number a but stays bigger than a " — think a = 2 and x = 2.1 , 2.01 , 2.001 , … The little raised + is a direction label , not a plus-sign to compute with.
x → a − means "x slides toward a staying smaller " — x = 1.9 , 1.99 , …
⌊ x ⌋ is the greatest integer ≤ x — you round down to the nearest whole number. ⌊ 3.7 ⌋ = 3 , ⌊ 3 ⌋ = 3 , ⌊ − 0.2 ⌋ = − 1 . (More in Greatest integer / floor function .)
The one law we keep using:
lim x → a f ( x ) = L ⟺ lim x → a − f ( x ) = lim x → a + f ( x ) = L .
Both sides must exist and be equal . If either fails, the two-sided limit does not exist.
The figure above is our map: a single number line with a marked. The coral arrow creeps in from the right (a + ), the lavender arrow from the left (a − ). Every example below is just "what value is each arrow's function heading toward?"
Worked example Ex 1 · Two rules that happen to meet
f ( x ) = { 2 x + 1 x 2 + 2 x ≤ 1 x > 1
Find lim x → 1 − f , lim x → 1 + f , and decide the two-sided limit.
Forecast: the left rule at x = 1 gives 3 ; the right rule gives 1 + 2 = 3 . Guess: they match, limit = 3 .
Left side: for x < 1 we use 2 x + 1 . Push x → 1 − : 2 ( 1 ) + 1 = 3 .
Why this step? On the left of 1 , only the rule 2 x + 1 is "active", and it is a plain line — safe to substitute the approached value.
Right side: for x > 1 we use x 2 + 2 . Push x → 1 + : 1 2 + 2 = 3 .
Why this step? On the right, the active rule switches to x 2 + 2 ; again a polynomial, so substitution is legal.
Combine: both sides give 3 , equal, so by the law lim x → 1 f = 3 .
Verify: test near-points. f ( 0.99 ) = 2 ( 0.99 ) + 1 = 2.98 ≈ 3 ✓. f ( 1.01 ) = ( 1.01 ) 2 + 2 = 3.0201 ≈ 3 ✓. Both hug 3 .
Worked example Ex 2 · The classic sign-jump
s ( x ) = x − 4 ∣ x − 4∣ , x = 4. Find both one-sided limits at 4.
Forecast: absolute value over the thing inside — this is a "+ 1 / − 1 " splitter. Guess: right = + 1 , left = − 1 , no two-sided limit.
Right side (x > 4 ): then x − 4 > 0 , so ∣ x − 4∣ = x − 4 . Ratio = x − 4 x − 4 = 1 .
Why this step? The definition of absolute value: for a positive quantity, ∣ u ∣ = u . Here u = x − 4 is positive when x > 4 .
So lim x → 4 + s = 1 .
Left side (x < 4 ): then x − 4 < 0 , so ∣ x − 4∣ = − ( x − 4 ) . Ratio = x − 4 − ( x − 4 ) = − 1 .
Why this step? For a negative quantity, ∣ u ∣ = − u (flip the sign to make it positive).
So lim x → 4 − s = − 1 . Since 1 = − 1 , the two-sided limit does not exist — a jump of size 2 .
Verify: s ( 4.5 ) = 0.5 ∣0.5∣ = 1 ✓ and s ( 3.5 ) = − 0.5 ∣ − 0.5∣ = − 0.5 0.5 = − 1 ✓.
Common mistake "Both sides exist (
± 1 ), so the limit exists."
The fix: existence on each side is only half the law — they must also be equal . Here they aren't. This is a jump discontinuity .
Worked example Ex 3 · A square-root wall
Find the one-sided limits of r ( x ) = x − 2 at x = 2 .
Forecast: you can't square-root a negative real. So the left side has nowhere to stand . Guess: only the right limit exists, and it's 0 .
Domain check first. x − 2 is real only when x − 2 ≥ 0 , i.e. x ≥ 2 .
Why this step? A limit can only "approach" through points where the function is defined . No domain on the left = no left-hand limit to speak of.
Right side (x > 2 ): as x → 2 + , x − 2 → 0 + , so x − 2 → 0 . Thus lim x → 2 + r = 0 .
Left side (x < 2 ): the function is undefined; lim x → 2 − r does not exist (no domain).
Two-sided: does not exist, because one side is missing — but for a purpose-built reason: it's a domain edge , not a mismatch.
Verify: r ( 2.01 ) = 0.01 = 0.1 , and r ( 2.0001 ) = 0.0001 = 0.01 — shrinking toward 0 ✓. r ( 1.9 ) = − 0.1 is not a real number ✓.
Worked example Ex 4 · Sides blow up in opposite directions
Find the one-sided limits of p ( x ) = x − 5 1 at x = 5 .
Forecast: tiny denominator = huge output; the sign of the denominator decides up or down. Guess: right → + ∞ , left → − ∞ .
Right side (x > 5 ): x − 5 is a tiny positive number, e.g. 0.001 . Then 0.001 1 = 1000 , growing without bound. So lim x → 5 + p = + ∞ .
Why this step? 1 divided by a shrinking positive number gets arbitrarily large and positive.
Left side (x < 5 ): x − 5 is a tiny negative number, e.g. − 0.001 . Then − 0.001 1 = − 1000 . So lim x → 5 − p = − ∞ .
Two-sided: does not exist — sides diverge to opposite infinities. This is a vertical asymptote at x = 5 .
Verify: p ( 5.001 ) = 0.001 1 = 1000 > 0 ✓, p ( 4.999 ) = − 0.001 1 = − 1000 < 0 ✓.
lim x → 5 + p = + ∞ , so this one-sided limit exists ."
The fix: + ∞ is not a real number . We say the limit diverges to + ∞ — it describes the runaway behaviour, but counts as a non-existent (infinite) limit.
Worked example Ex 5 · Floor where the approach point is NOT an integer
Find both one-sided limits of ⌊ 2 x ⌋ at x = 1.5 .
Forecast: floor only jumps where its input hits an integer. Input is 2 x ; at x = 1.5 that's 2 ( 1.5 ) = 3 , an integer — so a jump does happen here. Guess: left = 2 , right = 3 .
Find where the inside is integer. 2 x = 3 exactly at x = 1.5 . So the "step" of the floor lands precisely on our point.
Why this step? ⌊ ⋅ ⌋ is constant between integers and jumps at them; we must check whether 2 x crosses an integer at x = 1.5 . It does.
Left side (x < 1.5 ): say x = 1.49 , then 2 x = 2.98 , and ⌊ 2.98 ⌋ = 2 . So lim x → 1. 5 − ⌊ 2 x ⌋ = 2 .
Why this step? Just below 1.5 , 2 x is just below 3 , so it floors down to 2 .
Right side (x > 1.5 ): say x = 1.51 , then 2 x = 3.02 , and ⌊ 3.02 ⌋ = 3 . So lim x → 1. 5 + ⌊ 2 x ⌋ = 3 .
Two-sided: 2 = 3 , so it does not exist — a jump of 1 .
Verify: ⌊ 2 ( 1.499 )⌋ = ⌊ 2.998 ⌋ = 2 ✓ and ⌊ 2 ( 1.501 )⌋ = ⌊ 3.002 ⌋ = 3 ✓.
Worked example Ex 6 · You cannot just plug in
Find lim x → 3 − and lim x → 3 + of q ( x ) = x − 3 x 2 − 9 .
Forecast: at x = 3 we'd get 0 0 — undefined, so plugging in is illegal. But it factors. Guess: both sides give 6 ; limit exists = 6 .
Diagnose. Direct substitution gives 3 − 3 3 2 − 9 = 0 0 , an indeterminate form — no information yet.
Why this step? 0 0 means "cannot read off"; we must simplify before taking the limit.
Factor the top. x 2 − 9 = ( x − 3 ) ( x + 3 ) (difference of squares). So q ( x ) = x − 3 ( x − 3 ) ( x + 3 ) = x + 3 for x = 3 .
Why this step? Cancelling the shared ( x − 3 ) removes the trouble; valid because x = 3 near the limit (we never touch x = 3 ).
Now both sides use x + 3 . Left: x → 3 − , x + 3 → 6 . Right: x → 3 + , x + 3 → 6 .
Both give 6 , equal, so lim x → 3 q = 6 . (The hole at x = 3 is removable .)
Verify: q ( 2.99 ) = 2.99 − 3 2.9 9 2 − 9 = − 0.01 − 0.0599 = 5.99 ✓ and q ( 3.01 ) = 0.01 3.0 1 2 − 9 = 6.01 ✓.
Worked example Ex 7 · Infinite oscillation, tamed
Find lim x → 0 + and lim x → 0 − of w ( x ) = x sin ( x 1 ) .
Forecast: sin ( 1/ x ) oscillates crazily near 0 , but x shrinks to 0 . A small number times a bounded number → 0 . Guess: both sides = 0 .
Bound the wiggle. For every x = 0 , − 1 ≤ sin ( 1/ x ) ≤ 1 .
Why this step? sin of anything stays inside [ − 1 , 1 ] ; we don't need to know where it is, only that it's bounded.
Multiply by ∣ x ∣ . Then − ∣ x ∣ ≤ x sin ( 1/ x ) ≤ ∣ x ∣ .
Why this step? Multiplying the bound by ∣ x ∣ traps w ( x ) between two things that both go to 0 — the squeeze .
Both bounds → 0. As x → 0 from either side, ∣ x ∣ → 0 and − ∣ x ∣ → 0 , so w ( x ) is squeezed to 0 .
Therefore lim x → 0 − w = lim x → 0 + w = 0 , and the two-sided limit is 0 — even though w oscillates infinitely often.
Verify: w ( 0.01 ) = 0.01 sin ( 100 ) ≈ 0.01 ( − 0.506 ) = − 0.00506 , tiny ✓. w ( − 0.01 ) = − 0.01 sin ( − 100 ) ≈ − 0.00506 , also tiny ✓. Both near 0 .
Worked example Ex 8 · A parking-garage price jump
A garage charges $3 per hour, but the moment you pass 2 hours you are billed a flat $8 for the whole session up to 4 hours. Model the cost C ( t ) and find the one-sided limits at t = 2 hours.
Forecast: below 2 hours cost climbs toward 3\times2=\ 6; a t / a f t er 2 hours it jumps to \ 8. Guess: left = 6 , right = 8 , no two-sided limit — the customer sees a price cliff.
Translate to a piecewise rule.
C ( t ) = { 3 t 8 0 ≤ t < 2 2 ≤ t ≤ 4
Why this step? "Per-hour up to 2h" is the line 3 t ; "flat 8 after passing 2h" is a constant.
Left side (t → 2 − ): use 3 t : 3 ( 2 ) = 6 . So \lim_{t\to2^-}C=\ 6. ∗ W h y t hi ss t e p ? ∗ J u s t u n d er 2$ hours you're still on the per-hour meter.
Right side (t → 2 + ): use 8 : constant, so \lim_{t\to2^+}C=\ 8$.
Interpret: 6 = 8 , no two-sided limit — a real jump discontinuity of $2. That extra $2 is the "penalty" for crossing the 2-hour line.
Verify (units $): C ( 1.99 ) = 3 ( 1.99 ) = 5.97 ≈ 6 ✓; C ( 2.01 ) = 8 ✓. Prices behave as forecast.
Worked example Ex 9 · Make the pieces meet
f ( x ) = { k x + 2 x 2 + k x < 1 x ≥ 1
For what value of k does lim x → 1 f ( x ) exist? (This ties into Continuity at a point .)
Forecast: the two-sided limit exists only if the sides match; set them equal and solve for k . Guess: one clean value of k .
Left side (x → 1 − ): rule k x + 2 gives k ( 1 ) + 2 = k + 2 .
Why this step? Below 1 the active rule is the line k x + 2 ; substitute the approached value 1 .
Right side (x → 1 + ): rule x 2 + k gives 1 + k .
Force equality (the law). For the two-sided limit to exist:
k + 2 = 1 + k ⟹ 2 = 1.
Why this step? We demanded the sides match; but 2 = 1 is false for every k .
Conclusion: there is no value of k that makes the limit exist — the constant terms clash no matter what. (A sneaky exam trap: the unknown you're solving for cancels out.)
Verify: subtract the two expressions: ( k + 2 ) − ( 1 + k ) = 1 = 0 , independent of k ✓. The sides always differ by exactly 1 .
Common mistake "Set the sides equal, get
k = … , done."
The fix: always check whether your equation is solvable. Here the k 's cancel and you're left with a false statement — the honest answer is "no such k ."
Recall Which example handled which shape?
A agree ::: Ex 1 (limit = 3 )
B jump ::: Ex 2 (+ 1 vs − 1 )
C domain edge ::: Ex 3 (right = 0 , left undefined)
D pole ::: Ex 4 (+ ∞ vs − ∞ )
E floor twist ::: Ex 5 (2 vs 3 )
F removable 0/0 ::: Ex 6 (both = 6 )
G squeeze wiggle ::: Ex 7 (both = 0 )
H word problem ::: Ex 8 ($6 vs $8)
I unknown constant ::: Ex 9 (no k works)
Mnemonic The scan order for any one-sided problem
"DEFAULT DAP" — check D omain, then E valuate each side, then F orm-check (0/0 ? ∞ ?), A lign (are they equal?), P ronounce the verdict.
One-sided Limits — Left-hand & Right-hand (parent)
Limit of a function — intuitive & ε-δ definition
Continuity at a point
Jump, removable & infinite discontinuities
Vertical asymptotes
Greatest integer / floor function
Differentiability — left & right derivatives
Form check zero over zero or infinity