Gremlin Steps
Gremlin is a graph traversal language developed by TinkerPop. Faunus, as of the current release, provides a subset of the Gremlin steps. For sake of clarity, the Gremlin that compiles down to Pipes and is used for real-time traversals against a graph database is denoted Gremlin/Pipes. The Gremlin that is distributed with Faunus and compiles down to MapReduce is denoted Gremlin/Faunus.
There are major conceptual differences between Gremlin/Pipes and Gremlin/Faunus. These differences are itemized below.
- Gremlin/Faunus is breadth-first. Gremlin/Pipes is depth-first.
- Gremlin/Faunus processes rows of an adjacency matrix (vertex-centric). Gremlin/Pipes processes elements in a linked-list (element-centric).
- Gremlin/Faunus is inherently parallel with each vertex being operated on simultaneously. Gremlin/Pipes is inherently serial with one element at a time being pushed through the pipeline.
The table below presents the steps provided with Gremlin/Faunus.
| Gremlin/Faunus step | Description | Reduce? |
|---|---|---|
| transforms | ||
_() |
the identity step that simply updates Hadoop counters | false |
transform(closure) |
call the closure with current elements as argument | false |
V() |
start a traversal at all vertices | false |
E() |
start a traversal at all edges | false |
v(id..) |
start a traversal a some vertices | false |
out(labels..) |
traverse out from current vertices to label-related vertices | true |
in(labels..) |
traverse in from current vertices to label-related vertices | true |
outE(labels..) |
traverse from current vertices to outgoing label edges | true |
inE(labels..) |
traverse from current vertices to incoming label edges | true |
outV() |
traverse from current edges to outgoing vertices | false |
inV() |
traverse from current edges to incoming vertices | false |
property(key,class?) |
emit the the property value of current elements (specify class? for optimization) | false |
path() |
emit the path taken up to the current elements | false |
order(order,key) |
order the previous properties and identify element by key | true |
| filters | ||
filter(closure) |
A generic filter that takes a boolean-closure | false |
has(key,values..) |
filter elements that don’t have key/value property (values or’d) | false |
has(key,compare,values..) |
filter elements that don’t have compared key/value property (values or’d) | false |
hasNot(key,values..) |
filter elements that do have key/value property (values or’d) | false |
hasNot(key,compare,values..) |
filter elements that do have compared key/value property (values or’d) | false |
interval(key,value1,value2) |
filter elements that don’t have value in range | false |
dedup() |
remove any duplicates in current traversal step | false |
back(step) |
go back to elements seen previous | true |
| side-effects | ||
sideEffect(closure) |
mutate the current element according to the provided closure | false |
as(name) |
name the previous step for future reference | false |
linkOut(step,label,key?) |
project an outgoing edge to a previous vertex with provided key being an optional weight | true |
linkIn(step,label,key?) |
project an incoming edge to a previous vertex with provided key being an optional weight | true |
keep() |
remove all elements not at current step (of current type) | true/false |
drop() |
remove all elements at current step (of current type) | true/false |
groupCount() |
count the number of previously specified properties | true |
groupCount(closure,closure) |
process current elements with closure and incr count with closure | true |