Can an Eulerian circuit visit a vertex more than once?

Yes. Each edge is used once, but vertices can repeat.

Does a graph with an open Eulerian trail necessarily count as Eulerian here?

No. The NetworkX predicate in this entry requires a closed Eulerian circuit.

What if a cycle has an extra isolated node?

Under NetworkX's connectivity convention, the full graph does not pass is_eulerian.

AI Glossary / Graphs and symbolic AI / Intermediate

Eulerian

Definition

A graph is Eulerian when it has a circuit that traverses every edge exactly once and returns to its starting vertex. The route may visit a vertex several times; the restriction concerns edges. In NetworkX, is_eulerian checks this circuit property using connectivity and degree conditions appropriate to directed or undirected graphs.

What it means

The central question is whether every connection can be covered once in one closed route. That is different from visiting every location once, which is associated with Hamiltonian paths and cycles. An Eulerian circuit concerns the links, even when several links meet at the same location.
This distinction is useful in network inspection and symbolic graph exercises. A successful edge-covering circuit is a structural result. It does not automatically satisfy travel times, capacity limits, prohibited turns, or other conditions that a real route-planning problem might impose.

How it works

For a connected undirected graph with at least one edge, an Eulerian circuit exists exactly when every vertex has even degree. Arriving at a vertex and leaving it pairs up incident edges, including the starting vertex when the route closes.
For a directed graph in the relevant connected setting, each vertex must have equal in-degree and out-degree, with the required strong connectivity. NetworkX checks these conditions. Its treatment of disconnected graphs includes isolated vertices: adding an isolated vertex to an otherwise Eulerian component makes the full graph fail its connectivity test. Applications that intentionally ignore isolates must do so explicitly rather than assume the library ignores them.

Three practical examples

A graph tutorial tests whether every link in a small network can be traversed once while returning to the origin.
An inspection-planning prototype uses Eulerian structure as a starting point before adding operational restrictions and repeated travel where needed.
A directed workflow model checks whether balanced incoming and outgoing transitions support a closed traversal of all modeled transitions.

Think of it this way

Think of drawing every line in a connected shape without lifting the pen or tracing any line twice, finishing where you began. This captures the edge-once circuit idea. The analogy is limited because crossing lines are vertices only if the graph defines them as connected, and real movement may impose direction or turning restrictions absent from the drawing.

Worked example

Take an undirected cycle with vertices A, B, C, and D and edges AB, BC, CD, and DA. Every vertex has degree 2 and the graph is connected. The route A-B-C-D-A uses all four edges once and returns to A, so the graph is Eulerian.
Now add diagonal AC. Degrees of A and C become 3, while B and D remain 2. The graph no longer has an Eulerian circuit. It does have an open edge-once trail, for example A-B-C-D-A-C, which starts at A and ends at C.
These outcomes distinguish a circuit from an open trail. Finally, restore the four-edge cycle and add isolated vertex E. NetworkX's is_eulerian returns false for the full disconnected graph, even though the cycle component still has its own circuit. Whether E should be removed depends on what the graph is intended to represent.

Illustrative case study

In a fictional network workshop, the Maple team models service corridors as undirected edges. Its first check confirms that a cycle-shaped network has an Eulerian circuit. A participant then adds an isolated storage location to the node inventory and sees the result change.
The team reports the circuit for the connected corridor component separately and labels the isolated location as outside that component. It does not silently discard the location from the inventory. The invented case explains an implementation convention and a modeling decision, rather than claiming a complete solution to an actual inspection operation.

Common misconceptions and mistakes

Even degrees alone are insufficient when the graph has separate edge-bearing components. Each component might have a circuit, but one continuous route cannot move between disconnected components. In directed graphs, total degree being even is not a substitute for checking incoming and outgoing balance.
Do not confuse an Eulerian circuit with a shortest path, a Hamiltonian cycle, or a route that covers edges at least once. If extra traversals are allowed, a different optimization problem may be relevant. Parallel edges also remain distinct edges and must each be accounted for in a multigraph.

Business use

Eulerian checks can identify whether a simplified network admits exact once-only edge coverage. This is useful in early analysis, testing graph representations, and explaining why an additional traversal may be necessary.
Before applying the result, confirm what vertices and edges mean, whether directions are accurate, and how isolated locations are handled. If the operational goal permits repeats or requires additional constraints, formulate that goal separately. The yes-or-no Eulerian property is a foundation for reasoning about routes, not an all-purpose measure of route quality.

Check your understanding

Why does adding diagonal AC to the four-vertex cycle destroy the Eulerian circuit property?

Answer: It makes A and C odd-degree vertices. A connected undirected graph with exactly those two odd vertices can have an open Eulerian trail between them, but cannot have a closed edge-once circuit.

Related terms

Directed graph
Eulerian circuit
Eulerian trail
vertex degree
connected graph
strongly connected graph
Hamiltonian cycle
route inspection

Learn next

Learn next about constructing an Eulerian circuit. Trace one in a graph made of two cycles sharing a vertex, marking each edge as it is used. Compare the repeated vertex with the absence of repeated edges, then test what changes when one connection is removed.

Sources

NetworkX official is_eulerian documentation: https://networkx.org/documentation/stable/reference/algorithms/generated/networkx.algorithms.euler.is_eulerian.html NetworkX official Euler algorithms source: https://networkx.org/documentation/stable/_modules/networkx/algorithms/euler.html

Previous: Euler tour technique Next: Eulerian video magnification